SOMALI REFUGEE RESETTLEMENT AND LOCATIONAL DETERMINANTS IN THE
UNITED STATES
By
Jay L. Newberry

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Geography
2011

ABSTRACT
SOMALI REFUGEE RESETTLEMENT AND LOCATIONAL DETERMINANTS IN THE
UNITED STATES
By
Jay L. Newberry
The United States is a classic country of immigration, but it is also a traditional country
of resettlement for refugees. Since the enactment of the 1980 Refugee Act, the United States has
resettled over two and a half million refugees. Out of all of the groups that have been resettled,
there is one group that has gained national and even international attention as a result of their
migration practices here in the United States. In 2002 Lewiston, Maine became a site for the
secondary migration of Somali refugees. The financial hardships the town faced as a result of
that migration caused then-Mayor Laurier T. Raymond to write a public letter pleading for the
Somali elders to put a halt to the migration. The letter, which was construed as racist, attracted
national and international attention. At the time, Lewiston was not a resettlement city, thus it did
not have the resources or skills to deal with this new population. Therefore, the challenge and
cost to settle the Somali initially fell on the community. The events that took place in Lewiston
– the unexpected influx of Somali secondary migrants – have taken place in numerous cities and
towns across America. The problem is that little is known about the factors that influence the
residential distribution of the Somali or the factors that motivate them to embark on secondary
migrations after their initial resettlement in communities in the United States. Using both
qualitative and quantitative methods, this research analyzes several social, economic, and
demographic variables from various U.S. government databases in order to identify the order of
the factors most associated with not only the residential location of the Somalis, but also their
migration destinations. To further elaborate on the quantitative findings, this research employs

qualitative instruments such as surveys, unstructured interviews, and participant observation
conducted in Columbus, Ohio – home to the second largest concentration of Somalis in the
United States.

Results indicate that the Somali were dispersing from their initial resettlement
communities in secondary migrations not only to frontier communities, but also to communities
with an existing Somali population. The time frame of the data indicates that this was a period
of ethnic community building for the Somali. This fits in with the results of the analysis of
diversity, which indicated that the growing Somali communities became more diverse with
respect to clan-family representation over time. This growing diversity can be contributed to the
expansion of the number of Somali tribal/clan groups considered to be in imminent danger as the
civil conflict progressed. Survey results suggest that the presence of one‘s own clan would never
serve as the primary reason for secondary migration, but it would serve as a strong motivator
because of the social network resources it could provide at the destination. While the destination
communities for Somali secondary migrants were somewhat similar to the communities where
they originally resided – which typify refugee resettlement communities – subtle important
differences were noted.

Results of the principal components regression on migration

determinants revealed that these differences included, but were not limited to, the presence of an
older population, lower crime rates, and a decreased importance placed on the black or white
racial demographics. Results also indicate that, while the employment aspects of the migrant
destinations decreased in importance, the health of the welfare aspects associated with the
destination of the Somali migrants increased in importance.

Copyright by
JAY L. NEWBERRY
2011

ACKNOWLEDGEMENTS

The writing of this dissertation has constituted one of the most challenging academic endeavors I
have ever faced. With this said, I would like to give thanks to the following people, for without
their guidance, patience, and support, this research would have never been completed;
–

Dr. Joe T. Darden; Advisor, Dissertation Committee Chair, Mentor.

–

Dr. Steven J. Gold, Dr. Sue Grady, and Dr. Bruce Pigozzi; Dissertation Committee
Members and role models.

–

Karina Harty-Morrison and the staff at the Community Refugee and Immigration
Services (CRIS) in Columbus, Ohio.

–

The Somalis survey respondents in Columbus, Ohio.

v

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES .........................................................................................................................x
CHAPTER 1: INTRODUCTION ....................................................................................................1
Background ..................................................................................................................................1
United States Refugee Policy.......................................................................................................5
The Countries of Origin of Refugees ...........................................................................................7
Background to the Research Problem ..........................................................................................8
Purpose and Scope of the Study.................................................................................................12
Theoretical Framework ..............................................................................................................15
Ethnic Origins ........................................................................................................................15
Behavioral Approach to Migration........................................................................................18
Network Approach to Migration ............................................................................................19
Research Questions ....................................................................................................................20
Data and Methods ..................................................................................................................21
The Study Area .......................................................................................................................24
Significance of the Study ...........................................................................................................25
Summary of Remaining Chapters ..............................................................................................26
CHAPTER 2: THE SOMALI ........................................................................................................27
Somalia ......................................................................................................................................27
Geography and Climate of Somalia.......................................................................................27
The Somali People .................................................................................................................29
The Colonization of Somalia .....................................................................................................35
The Aftermath of Colonization ...............................................................................................37
Post-Independence: The Siyaad Barre Years .............................................................................38
Somali Nationalism ................................................................................................................40
The Somali Conflict ...................................................................................................................41
Barriers to Unification...........................................................................................................43
Somali 4.5 System ..................................................................................................................45
Somali Refugee Resettlement in the United States....................................................................47
Corruption in the Refugee Priority 3 Program ...........................................................................52
American Concerns about Somalis ............................................................................................53
Somali Americans Linked to Terrorism .................................................................................54
Somali Pirates ........................................................................................................................55
Chapter Summary ......................................................................................................................56
CHAPTER 3: REFUGEE AND REFUGEE RESETTLEMENT ..................................................58
Refugees .....................................................................................................................................58
Pre-1980 Refugee Legislation....................................................................................................59
The Refugee Act of 1980 ...........................................................................................................60

vi

Refugee Resettlement ................................................................................................................62
United States Refugee Policy .................................................................................................62
Changes in Refugee Policy ....................................................................................................68
Refugee Model and a Refugee Type .......................................................................................70
Voluntary Agencies (VOLAGS)................................................................................................71
The VOLAG-Government Partnership ..................................................................................73
Problems in Refugee Resettlement ............................................................................................74
The Welfare System ................................................................................................................75
Refugees and the Welfare Magnet States ..................................................................................77
The Personal Responsibility and Work Opportunity Reconciliation Act...............................78
Chapter Summary ......................................................................................................................79
CHAPTER 4: REVIEW OF THE LITERATURE ........................................................................82
Studies of Somali Adaption .......................................................................................................83
Immigrant Geographic Distributions .........................................................................................85
Studies of Locational Determinants .......................................................................................85
Studies Specific to Refugees ...................................................................................................90
Studies Specific to the Somali ................................................................................................94
Gaps in the Literature.................................................................................................................95
CHAPTER 5: DATA AND METHODS ......................................................................................98
Step 1: Determination of Clan-Family in Cities of Initial Resettlement ..................................98
Step 2: Determination of Somali Dispersion ..........................................................................101
Step 3: Identification of Somali Locational Determinants .....................................................103
Step 4: Identify Somali Migration Determinants ....................................................................106
Participant Observation.......................................................................................................111
Snowball Sampling...............................................................................................................112
Columbus, Ohio as Case ......................................................................................................116
Limitations of the Data ............................................................................................................122
CHAPTER 6: RESEARCH FINDINGS .....................................................................................124
Determination of Clan-Family Diversity in Cities of Initial Resettlement ..............................124
Determination of Somali Dispersion .......................................................................................125
Identification of Somali Locational Determinants ...................................................................126
Identification of Somali Migration Determinants ....................................................................135
Determination of Cities Associated with Specific Clans .........................................................143
Organization Surveys ...........................................................................................................144
Somali Migrant Surveys .......................................................................................................146
CHAPTER 7: SUMMARY AND CONCLUSION ....................................................................153
APPENDIX ..................................................................................................................................161
BIBLIOGRAPHY ........................................................................................................................173

vii

LIST OF TABLES

Table 1.1: Census 2000 special tabulation – foreign-born residency. ......................................... 10
Table 2.1: Somali political factions 1964 - 1969. ........................................................................ 39
Table 2.2: Top five refugee group admissions to the United States 1997 - 2009. ....................... 42
Table 2.3: Somali admissions by category 1990 - 2009. ............................................................. 49
Table 2.4: Resettlement figures by clan-family membership 1990 - 2009. ................................. 52
Table 3.1: Refugee Admissions and Ceilings 1990 - 2009. ......................................................... 67
Table 3.2: Top ten refugee resettlement states and totals 1983 - 2004. ....................................... 68
Table 3.3: Refugee eligibility timeline for major federal assistance programs. .......................... 69
Table 3.4: Refugees LPR by region of origin 1946 - 2000. ......................................................... 71
Table 3.5: Refugee resettlement voluntary agencies. .................................................................. 72
Table 5.1: Variables used for step 3 - locational determinants. .................................................. 104
Table 5.2: Variables used for step 4 - migration determinants. .................................................. 107
Table 5.3: ACS demographic estimates - 2005-2009 for Columbus, OH. ................................. 117
Table 5.4: Top 10 Somali groups resettled in Columbus, OH - 1990 - 2009. ............................ 119
Table 6.1: Resettlement city diversity (N=179). ........................................................................ 124
Table 6.2: Index of relative clustering. ...................................................................................... 125
Table 6.3: PCA Locational determinants - dimensions, loadings, and variance (n=1667). ...... 127
Table 6.4: Regression on locational dimension scores. ............................................................. 130
Table 6.5: PCR results - locational determinants....................................................................... 131
Table 6.6: PCA Migration determinants - dimensions, loadings, and variance (N=1667). ....... 136
Table 6.7: Regression on migration dimension scores. ............................................................. 139

viii

Table 6.8: PCR results - migration determinants. ...................................................................... 141
Table 6.9: Organization survey results. ..................................................................................... 145
Table 6.10: Somali respondent background (n=48). .................................................................. 147
Table 6.11: Reason for migrating. ............................................................................................. 147
Table 6.12: Influences on the decision to migrate. .................................................................... 150
Table 6.13: Support in migration. .............................................................................................. 151
Table A.1: Partial listing of the Darood clan-family ................................................................. 162
Table A.2: Partial listing of Dir, Isaaq, Hawiye, and Rahanweyn clan-families. ...................... 163
Table A.3: Variables for locational determinants. ..................................................................... 164
Table A.4: Variables for migration determinants. ..................................................................... 165
Table A.5: Survey for Columbus Somali................................................................................... 166
Table A.6: Somali Organization Survey Questions ................................................................... 169

ix

LIST OF FIGURES

Figure 2.1: Geographic location of Somalia. For interpretation of the references to color in this
and all other figures, the reader is referred to the electronic version of this dissertation. ............ 28
Figure 2.2: Basic structure of Somali segmentary lineage. ......................................................... 31
Figure 2.3: Map of the distribution of Somali clan-families in Somalia. .................................... 33
Figure 2.4: Colonial partitioning of Somali territory mid-19th century. ..................................... 36
Figure 2.5: Percentage of refugees from the African region who are Somali. ............................ 43
Figure 2.6: Somali admissions 1990 - 2009................................................................................. 48
Figure 2.7: Somali distribution according to the ORR initial community of resettlement. ......... 51
Figure 3.1: The 22 countries actively participating in the resettlement of refugee. .................... 63
Figure 3.2: Graphic depiction of refugee admission 1980 - 2009. .............................................. 64
Figure 3.3: Refugee admissions and ceilings. .............................................................................. 65
Figure 5.1: Forms of snowball sampling. .................................................................................. 114
Figure 5.2: Somali ORR resettlement in Columbus by year. .................................................... 119
Figure 5.3: Somali residential distribution in Columbus. .......................................................... 121

x

CHAPTER 1: INTRODUCTION

Background
The Somali are a relatively new and growing group of immigrants in the United States
owing to civil war between rival clans and political entities that left Somalia in ruins. Data
compiled from the Office of Refugee Resettlement (ORR) and the Department of Homeland
Security1 (DHS) reveals that over 108,000 Somali refugees, asylees, and I-942 non-immigrants
entered the United States between 1990 and 2009 at an average close to 5,414 per year over the
19-year period. Like most refugees who have been resettled in the United States, it was expected
there would be some secondary migration to established ethnic communities; however, in the
case of the Somali3, migration after they have initially been resettled in an American community
appears to occur frequently and sometimes on a large scale. These migrations – commonly
referred to as secondary migrations – and some of their negative consequences are evident in
media reports from places such as Lewiston, Maine. Somali secondary migrations across the
United States have been drawing media attention since the turn of the century.

Lewiston

garnered the most widespread media attention in 2001 when an influx of Somali secondary
migrants unexpectedly overwhelmed the city. The Somali secondary migration to Lewiston
gained national and even international attention when then-Mayor Laurier T. Raymond issued a

1

Archival data from the Department of Homeland Security can be found at
http://www.dhs.gov/files/statistics/publications/archive.shtm and the data from the Office of
Refugee Resettlement can be found at http://www.acf.hhs.gov/programs/orr/data/index.htm
2
I-94 refers an enumeration category of non-immigrants that are temporarily in the United States
and require a visa. This category includes – but is not exclusive to - vacationers, diplomats, and
students.
3
In this study, the Somali are categorized as refugees in general. While not all Somali in the U.S.
are refugees, nearly 94.5 percent of the Somalis legally admitted to the U.S. were done so as
refugees and asylees.
1

public plea in an open letter to the Somali community leaders urging them to discourage further
Somali secondary migration to Lewiston.

Refugee policy for the United States is made at the federal level, but local entities –
particularly voluntary agencies (VOLAGS) and state refugee coordinators – are responsible for
where refugees are resettled (Singer and Wilson 2006). When refugees are resettled in the
United States, placement is based on several criteria that include housing availability, the job
market, the presence of family, and community resources. These criteria are in place to ensure
that communities receiving the refugees have the ability to provide the basic necessities for their
newest members. Refugee policy does not contain any mandates or criteria specifying that
refugees must stay in their initial resettlement communities4, so many embark on secondary
migrations for locations with established ethnic communities where they expect to receive social
support from other compatriots. The Somali are no different in this respect. Many Somalis will
embark on secondary migrations to established Somali communities, which are typically
resettlement communities – i.e., ORR designated communities with the necessary resources,
government backing and experience to work with Somali refugees. On the other hand, many
Somalis will also make secondary migrations to frontier communities. In the context of this
dissertation, frontier communities refer to communities that initially had no Somali presence or
were not official resettlement communities – like Lewiston, Maine. An assessment of media
reports suggest that the residents of these communities are usually taken by surprise by the
swiftness and numbers in which the Somali arrive. When refugees are resettled in the United
States, the process requires financial as well as human resources for endeavors ranging from the
4

For the purpose of this dissertation, initial resettlement community refers to the community a
refugee is assigned to by the ORR once they have been granted entry into the United States.
2

provision of cash assistance and translation services to housing, education, and job placement.
These endeavors are normally carried out by VOLAGS. If the community is not an ‗official‘
resettlement community – as recognized by the ORR – and thus not receiving funding or support
from the ORR, then the endeavors typically handled by the VOLAGS fall on the community.
The sentiment of many communities who find themselves facing this situation where they are
suddenly charged with the responsibility of caring for a population that is not only new to the
community, but also new to the country, is aptly captured in the following passage from a
Church World Service interview with Fabian Talamante – Director of the Dallas-based Refugee
Services of Texas sub-office in Amarillo. Talamante‘s response in the following passage is to a
question pertaining to a recent influx of refugee secondary migrants (to include the Somali) into
Dumas, Texas:
―So what are we supposed to do, not help?‖ asked Talamante. ―Of course we help! I
wish we could get funding, but we still do the work‖ of making sure the secondary
migrants get the help they need. (Church World Services 2009)
The response provided by Talamante reflects a common reaction characteristic of many
communities. Thus, even though non-resettlement communities5 are under no obligation to help
secondary migrants, they do so anyway. The reason why these communities choose to extend the
help, however, is beyond the scope of this research.

Refugee secondary migrations to both resettlement and non-resettlement communities
can in some cases present problems. According to Nadeau (2003), the relocation to resettlement
and non-resettlement communities can generate fiscal, social, and program changes that require
amendments to local and statewide policy.

When refugees emigrate from their initial

5

The term ‗non-resettlement community‘ is used in this dissertation to refer to communities that
have not been officially recognized or designated as a resettlement community by the ORR.

3

resettlement community, the resettlement benefits endowed to them – social services, cash and
medical assistance – do not follow through to their new destination. Thus, when secondary
migrants arrive in non-resettlement communities where there are no programs in place to handle
the tasks associated with resettlement, services extended to help typically become public charges.
When there are numerous secondary migrants, the public charges become taxing on the
community as with the Somali secondary migrants in the Lewiston case. Similar issues can also
arise when a substantial amount of secondary migrants arrive in communities designated by the
ORR as ―official‖ resettlement communities. The money provided to resettlement communities
by the ORR is based on calculations for the set number of refugees sent to those communities.
Some resettlement communities have reserve funds for unexpected arrivals; however, when they
are inundated with secondary migrants, the reserve funds are exhausted and the services provided
for those additional unexpected arrivals become public charges as with the Somali secondary
migrants in the Columbus case. In Columbus, Ohio, over 300 Somalis migrated to the city
unexpectedly; and because the city was un-prepared, the families spent weeks living in homeless
shelters that were originally structured to house people for only a few days (Ohio Refugee
Services 2006). In Columbus, Lewiston, and many of the other communities that receive Somali
secondary migrants, the main issues raised concern the estimation of exactly how many Somalis
will be migrating to their location, why are they migrating to their location, and how long will
they stay. Despite the fact that Somalis have been coming to this country in significant numbers
for the past two decades, little is known about this population, – especially about their
motivations to migrate within the continental United States after they are assigned to a
resettlement community.

4

Somali secondary migration has been attributed to their previous lifestyle as ―nomads‖ in
Somalia trying to find ―greener pastures‖ (Horst 2007). Indeed, it has been noted that the Somali
population has been characterized as continuously moving – even after their resettlement in the
global diaspora (Kapteijns and Arman 2004). The nomadic explanation for Somali migration,
however, neglects to explain any role of clan culture – the culture that formed the basis of their
society for hundreds of years and may play a role in the secondary migration of Somalis
throughout America. In addition, the ‗greener pastures‘ analogy lacks definition and quantifiable
attributes. In Somalia, the nomadic way of life is associated with clans and clans are associated
with support and security; thus, the pertinent questions are – when Somalis migrate within the
United States following their initial resettlement into a community, is the decision individualistic
or is the decision influenced by clan allegiance? Is the spatial distribution of Somali in the
United States random, or is there a pattern to their distribution? A sizeable proportion of the
research that has examined refugee migration within the United States typically looked at
refugees as a whole; this generalizing approach, however, neglects the unique attributes of the
groups that make up the refugee population and in the process obscures any differences. This
type of aggregation commonly results in expectations concerning refugee distribution and
migration patterns, which may or may not be true once the refugees are disaggregated into their
individual ethnic or racial group. By specifically focusing on the Somali (i.e. a single refugee
group), this research corrects such over-sight.

United States Refugee Policy
Refugees are defined as persons who, owing to a well founded fear of persecution for
reasons of race, religion, nationality, membership of a particular social group or political

5

opinion, is outside the country of his/her nationality and is unable or, owing to such fear, is
unwilling to return (Hein 1993). There are over 22 resettlement countries, and 9 of these are
considered to be traditional6. Traditional resettlement countries are those who have historically
and consistently participated in the resettlement of refugees; the United States is by far the
largest of the resettlement countries accepting more refugees for resettlement than the other 21
countries combined (Patrick 2004).

As reflected in the volume of refugees accepted for

admission, an important goal of U.S. foreign policy is the continued assistance to and
resettlement of refugees; and since the 1980s, more than 2.5 million refugees have been resettled
in the United States. Refugee resettlement in the United States is largely based on humanitarian
concerns and the program itself is governed by the Refugee Act of 1980. The 1980 Refugee Act
introduced into American law a definition of refugee that was comparable to that established by
the United Nations (U.N.). Congress passed the Refugee Act of 1980 to provide a uniform
procedure by which to admit refugees and authorize federal assistance for their resettlement and
procurement of self-sufficiency (Bruno 2009). Policies stemming from the act also dictate the
dispersal of refugees; but, despite the fact that refugees have been dispersed throughout the
United States, 10 states in particular received the vast majority. In a comprehensive examination
of refugee resettlement conducted by Singer and Wilson (2006), it was discovered that California
headed a list of ten states that received nearly 70 percent of the 1.6 million refugees that had
been resettled between 1983 and 2004. The preceding figures are based on initial resettlement
counts; however, many refugees embark on secondary migrations soon after they arrive. Much
of the secondary migration occurs within the first few years after arrival within the United States

6

Traditional resettlement countries include: New Zealand, Norway, Sweden, United States,
Australia, Canada, Denmark, Finland, and the Netherlands. Switzerland used to be a traditional
resettlement country until it ended its resettlement program in 2002.
6

and then the refugee group becomes relatively stabilized in their geographic distribution after an
initial period of adjustment to life in the United States (Office of Refugee Resettlement 2006). It
has been suggested that better employment opportunities, established ethnic communities, better
welfare benefits, family reunification, or more congenial climates serve as factors that lure
refugees away from their initial resettlement communities (Office of Refugee Resettlement
2008) .

The Countries of Origin of Refugees
Traditionally, refugee groups entering the United States have been from developed
countries where they obtained skills that were transferable in a compatible culture; i.e., skills that
could be incorporated into a new area with a similar culture requiring few new resources. Since
the early 1990s, a steadily growing proportion of the refugees have been coming from
underdeveloped parts of the world; in many such instances, transferable skills and cultural
similarities are basically non-existent. These contemporary refugees require a greater amount of
assistance because it is typically harder for them to adapt to life in America given the substantial
difference between their culture and mainstream American culture.

The vast majority of

refugees are still coming from places such as Europe and Asia, but a growing number are coming
from the African region. The group for study in this research – the Somali – is from the African
region and has for the larger part constituted a stateless nation7 since 1991. Somalis, who are
predominately Muslim, come from a clan-based, primarily nomadic, African society. In the
United States, the Somali comprise a minority culture, religion, and race – this combination
provides for a different type of immigration issue (Schaid and Grossman 2007). Because of the
7

A stateless nation refers to a people without a state – state being an area that has a government,
a defined territory, and a permanent population (Getis, et al. 2011).
7

apparent uniqueness of this group – in addition to the fact that relatively little is known about
them – it is imperative that we understand their situation given that the social and political
turmoil in Somalia will continue to generate refugees for years to come (Goza 2007).

Background to the Research Problem
The Somali are highly mobile and it is reported that they are dispersing from their initial
resettlement communities to smaller cities and towns – this can present a problem because the
destination communities are often ill-informed about the size and make-up of this migrating
group. The number of Somalis arriving in many of the newly established migration streams
seems to be of priority concern, not only to the citizens at the destination communities, but also
to the leaders of those communities. Concern within a community constitutes an issue because
of the importance of the community. The importance of the community stems from the notion
that the community can provide direct assistance to refugees and that the community‘s response
can influence the government‘s policies concerning refugees (Jacobsen 1996). The following are
just three of an uncounted number of examples where the influx of Somali secondary migrants
has caused concern within destination communities:
Emporia, KS: In November 2007, citizens of Emporia believed their town would become
a major destination community for secondary migrants because within a matter of
months, somewhere from 750 to 1000 Somalis had migrated from various parts of the
United States to this small Kansas town. As local, state, and private leaders prepared,
they held meetings to expand on how they would spend the $154,000 emergency grant
earmarked for services, personnel, and space need to help settle the Somali; this
personnel would include outreach workers to help the Somalis with endeavors ranging
from applying for benefits, collection of missing documents, and affidavits of
relationship to bring others from the refugee camps in Africa to applying for housing,
counseling, and health services (Mlynar 2007). In less than a year, the majority of the
Somali community had decided to leave following the decision of Tyson Foods to close
its operations. (Berlin 2008)

8

Columbus, OH: Columbus is home to an estimated 30,000 Somali. During the first half
of 2005, over 300 Somali Bantu migrated to Columbus catching the city by surprise.
Over a dozen Somali families were living in a homeless shelter – originally designed to
house people for a few weeks – for months. Franklin County received over $650,000 in
federal grants to help with issues such as finding the Somalis jobs, enrolling their
children in school, enrolling the adults in English learning classes, housing, healthcare,
and other settlement issues (Juliano 2005). City and government social workers sought to
find more grants for emergency relief. It was reported that the new Somali migrants
arrived without any direct connection or job prospects in Columbus. Technical assistants
from the National Somali Bantu Project who went to Columbus to assess the situation
determined that the migrants felt desperate in their original resettlement communities
fearing that they would become homeless because they lacked jobs that paid enough to
cover food and housing for their families. The assistants concluded that misinformation
about jobs and housing availability in Columbus encouraged the migration (Ohio
Refugee Services 2006). The migrants came from over a dozen cities across the country
with Memphis TN, Atlanta GA, and Hartford CT accounting for the majority.
Lewiston, ME: Lewiston was the most publicized case of Somali internal migration. At
the beginning of 2001, no Somalis were living in Lewiston, but by the end of the year,
over 1,000 had migrated to this town with a population of 35,000. The Somali welfare
caseworker in Lewiston expected that close to 1,000 more Somalis would migrate to
Lewiston during the summer months of 2002 (Bouchard 2002). The Somali secondary
migrants arrived from communities all over the United States, but the majority came
directly from Atlanta, Georgia. Those that left Atlanta did so because of the over-taxed
welfare system, a fragile relation with the area‘s African American population, and a
desire to protect their children from assimilating into the American culture too quickly.
According to Tilove (2002), the Somali were seeking cheaper housing and a better
welfare system – by moving from Georgia to Maine, the Somali traded one of the least
generous welfare systems for one that is among the most generous. The Somali were also
seeking better schools and a better place to raise their children with ―fewer perils and
temptations‖ (Tilove 2002). The vast majority of the migrants had no money, no jobs,
and no job prospects when they arrived. City officials estimated that only 10 percent of
the Somali adults living in Lewiston were working because the local labor market had
little to offer unskilled workers without a high school diploma and the ability to speak
little or no English. Lewiston‘s welfare budget had doubled because most of the Somali
were seeking rent and food assistance (Bouchard 2002). The town received $200,000 in
federal grant assistance but more was needed. The grant was actually $50,000 less than
the amount the city spent on staff positions and programs created to incorporate Somali
students into the school system. Motivated by the escalating conditions, on October 1,
2002, Mayor Raymond issued a public letter to the Somali community urging restraint in
their migration practices because of the stress it placed on the city‘s finances and the
generosity of the residents. Mayor Raymond‘s letter drew sharp criticism from news
media and human rights activists from all over the country. (Bouchard 2002)

9

The frequent movement of Somalis is attributed to their nomadic heritage; however, this
alone does not explain the numbers that are seen moving to specific areas. Table 1.1 contains
data compiled from a U.S. Census 2000 special tabulation of foreign-born population
demographic and social characteristics for the United States. The table reveals that over 60
percent of the 34,815 Somalis (age 5 years and older) counted in the 2000 Census were still
living outside the United States in 1995. However, out of the 13,840 Somalis that were living in
the United States, only 3,650 (10.5 percent) were living in the same house in 2000 as they did in
1995. That percentage is much less than the native-born population – 55.7 percent of the
231,666,090 native-born people (age 5 and older) lived in the same house in 1995 as they did in
2000. This speaks to the sedentary nature of the populations.

TABLE 1. 1: Census 2000 special tabulation – foreign-born residency.
Group:
Population 5 years & over:
RESIDENCE IN 1995
Same house:
Different house in the U.S.
Same county
Different county
Same state
Different state
Elsewhere in 1995

Native-born

Somali

Vietnamese

Cuban

231,666,090

34,815

981,215

868,995

128,946,395
100,849,170
57,530,090
43,319,080
23,294,650
20,024,430
1,870,525

3,650
10,190
4,970
5,220
1,715
3,505
20,980

425,745
459,165
311,285
147,880
82,695
65,185
96,300

451,695
312,780
236,300
76,480
40,460
36,025
104,515

Source: (FILES: Census 2000 Special Tabulation (STP-159) - United States / prepared by
the U.S. Census Bureau 2002). Table FBP-1. Profile of Selected Demographic and Social
Characteristics: 2000.

In reference to the mobility of the populations, Table 1.1 also reveals that when the
native-born move, the destination is commonly within the state; however, when Somalis move,

10

the destination is typically in another state. Out of the 10,190 Somalis that lived in a different
house in 1995, 34.4 percent reported that the house was in different a state. This is compared to
only 19.9 percent for the native-born population. The percentage of Somalis that were living in
different states in 1995 than they did in 2000 is also higher when compared to other groups with
large refugee populations such as the Cubans and the Vietnamese. The percentage of the Cuban
and Vietnamese population who lived in a different state in 1995 than they did in 2000 was 11.5
and 14.2 percent respectively. Table 1.1 lends credence to the notion that the Somali are highly
mobile, and the Somali do have a nomadic heritage; however, that heritage is interlinked with
clan affiliation for safety and security – the act itself was primarily driven by environmental
factors:
War and feud occur frequently in Somaliland and constitute a further hazard to the
pastoralist in their movements. In an arid environment in which overgrazing is general
and where the human and stock populations press heavily upon the sparse grazing
resources available there is constant competition for access to pasture and water and
frequent lineage strife. This is most acute in the dry seasons but is generally
characteristic of the pastoral life as a whole. (Lewis 1994, 28)
Thus, continuous migration was an integral practice in Somalia because the survival of the
majority of the population depended on it. In fact, the Somali occupancy of the Horn of Africa
was the result of migration. According to Alexander (2001), the ancestral origin of the Somali
people was in the highlands of southern Ethiopia. Following a north-to-south route in search of
water and pastureland, the Somalis reached the southern shores of the Red Sea by the first
century A.D. and had occupied the entire ―Horn‖ by 100 A.D. (p.31).

The problem is that little is known about the Somali and their motivations for secondary
and subsequent migrations within the United States. In Somalia, as it has been made clear,
migrations were linked to survival; but what are the motivations when the natural environment is

11

no longer an opponent in the struggle to survive. Secondly, why are the secondary migrations so
large? In each of the cases outlined earlier, the destination communities were facing large
influxes of Somali over a short period of time.

If clan rather than the aspirations of the

individual influenced these secondary migrations, then this would begin to explain the size and
swiftness with which the Somali migrate in the United States. By better understanding the
motivation and organization – if any – behind Somali secondary and subsequent migrations,
destination communities can be better informed of what to expect and how to prepare for this
population. By being better prepared, destination communities can avoid some of the problems –
such as the financial issues – experienced by communities such as Columbus, Lewiston, and
Emporia.

Purpose and Scope of the Study
The overall purpose of this research is to extend the knowledge gained in the study of the
Somali global diaspora by focusing on the spatial aspects of the Somali in the United States. A
good example would be identifying common attributes in their migration destinations. The
current trend for studying the Somali compare globally dispersed Somalis by focusing on
questions regarding how cultural, political, economic, social, and racial context inform Somali
identity formations and opportunity structures (Kusow and Bjork 2007). While these aspects are
very important, the group‘s spatial distribution within their host society is just as critical because
opportunity is heavily influenced by location (Darden, Bagaka‘s and Ji 1997). In addition,
culture, politics, the economy, and the social and racial climate of one‘s surroundings are subject
to change dependent on location.

12

The processes of migration results in consequence – either negative or positive – to both
the origin and the destination communities. Population movements are generally seen as having
a negative effect on the originating community and a positive effect on the destination
community; but, in some instances, the migration can be burdensome on the destination
community (Clark 1986). The probability of burden is high when dealing with refugees because
this population typically has a lower rate of labor force participation; as refugees, some may be
traumatized and unprepared for employment while others – because of their age – are simply not
productive workers (Gold 1992).

It is generally accepted that refugees – after moving in

secondary migrations – tend to cluster (to gather together) in particular locations despite
resettlement policies of dispersion (Black 2001).

There is a possibility that Somalis are

dispersing (spreading out or scattering) and clustering – dispersing from their initial resettlement
communities and clustering in new communities by clan. If this is true, this ‗reshuffling‘ can be
detrimental in some situations. The U.S. policy of refugee dispersion is meant to prevent strain
on communities with respect to jobs, housing, and social services. Communities identified for
refugee resettlement by the ORR are based on a select set of criteria and are backed by an
allotted amount of government funding. If the refugees are not self-sufficient in the new clusters
formed by secondary and subsequent migrations, the negative impact – with respect to federal
monies – could be multiplied in that more communities outside designated resettlement
communities are affected. Somalis are of particular concern given the social constructs of race,
religion, and culture, which often put them at odds with the majority of American society. These
odds are compounded by linguistic and literacy barriers – in addition to health and psychological
issues. Given the perceived differences, Somalis are believed to be more difficult to resettle –
requiring more time, more funds, and more assistance for many needs. It is this belief that

13

facilitated former Senator Sam Brownback‘s rejection of the resettlement of Somali Bantu in
Kansas (Barnett 2003).

The influence of clan can be found throughout Somali society, and if the traditional ways
of the Somali are being reproduced here, then – to a lesser extent – the factors that facilitate clan
dependence are present here and the Somalis are simply relying on historic traditions to adapt. If
this is the case, then resettlement policies are compounding the situation by treating the Somalis
as one ethnic group when they are actually several distinct ethnic groups, and the secondary
migration is simply a way to reconstitute clan cohesion for safety and security. The clan is a
vital part of Somali history and clan conflict over resources was one of the primary factors
facilitating the Somali diaspora (Besteman 1999).
associated with specific regions of the country.

In Somalia clans are, to some extent,
According to Brons (2001), the original

territorial distribution of Somali clans is no longer distinct like it was at the end of the nineteenth
century, but in contemporary political settings, clans do claim specific territories – these claims
are predicated on traditional rights of access and control. Territoriality is thus a part of Somali
history and the association of specific clans with specific regions continues today in Somalia.
Therefore, another underlying purpose of this research is to determine if these aspects of Somali
society reconstituted in the global diaspora within the United States. If clan cohesion is
reconstituted in the secondary migrations of the Somali, there is the possibility that the territorial
aspect is also reconstituted.

The scope of this research is limited to the identification and examination of factors
influencing Somali secondary migration. This research uses several social, economic, and

14

demographic variables; the variables themselves are used in the context of identifying factors
associated with Somali secondary migration – i.e. descriptors of spatial processes. There is a
possibility that the analysis of these variables might highlight factors underlying inequalities in
the spatial location of the Somali; however, addressing these inequalities is beyond the scope of
this research. In addition, Somali secondary migrations are heavily influenced by the destination
communities; the motivations for the communities to accept or rejected the Somali secondary
migrants, however, is also beyond the scope of this research. The importance of this research is
that it will contribute to a foundation that will open the door for future research that will
eventually address any inequalities found in the spatial location of the Somalis.

Theoretical Framework
Ethnic Origins
The theoretical basis for this research is Jeremy Hein‘s ―ethnic-origins‖. According to
Hein (2006), there is mounting evidence that the history, culture, and politics of an immigrant
group shapes the way the immigrant group adapts to certain aspects of American society.
Culture refers to an entire way of life of a group of people; it is the values, beliefs, practices, and
behaviors that define and differentiate one group from another (Frazier 2006).

For the Somali, their culture is that of segmentary lineage or kinship8. In Somali society,
kinship – to some degree – determines who to marry, where to live, how to raise children, how to
8

According to Lewis (1982), the segmentary Somali lineage system has five major divisions;
clan-family, clan, sub-clan, primary lineage, and dia-paying group. This is a hierarchical
division with the clan-family at the upper limits of clanship. The clan represents the upper limit
of political action and has territorial properties. The primary lineage is the exogamous political
unit and the dia-paying group is the most basic political unit.
15

solve disputes, and more (Holly 1996). The foundation of the kinship system is the social
contract (xeer); some believe close group bonds formed not because of shared ancestry, but
because of explicit treaties defining the terms of their collective unity and regulating their
behavior as individuals via a system of delicts (Mohamed 2007). This implies cultural norms –
from the social contract – that have been instilled over centuries. These are cultural norms that
would not simply dissipate through transplantation to another country – especially in the limited
amount of time that the Somali have been here in the United States. Thus, utilizing ethnicorigins as a basis to identify Somali adaptation methods proves ideal because their actions can be
traced back to certain aspects of their culture and the social contract.

The purpose of ―ethnic-origins‖ – as conceived by Hein (2006) – is to explain the
differences seen in the amount of time it takes for immigrants to adapt to life in America and the
differences are explained by looking at the group‘s history in their country of origin. Hein used
Cambodian and Hmong refugees in his analysis of how the history, politics, and culture they
bring with them shape the way they adapt to American race and ethnic relations. These two
populations were chosen because both are subsumed within the same American racial
categorization as Asian, Southeast Asian, or ‗Other Asian.‘ Both groups entered at relatively the
same time and under the same venue as refugees. In addition, both groups occupied the same
low position on the American class hierarchy upon arrival in the United States; class background
is considered one of the more important factors shaping the adaption of contemporary
immigrants (p.37). Despite these remarkable similarities on American soil, the Cambodian and
Hmong history, politics, and culture are vastly different. The ethnic-origins hypothesis posits
that ―immigrant groups with distinct ethnic origins will have different forms of racial and ethnic

16

adaptation, even when these groups are similar in physical appearance, mode of incorporation,
and socioeconomic status‖ (p.34). In effect, immigrant groups with different histories will apply
different strategies for coping with the American assimilation process. This will occur regardless
of how similar the groups are biologically, how similar the situation that led to their entry into
the United States, or how similar their reception upon entry into the United States. Hein
stipulates that concepts nurtured in the country of origin pertaining to societal membership such
as religious values and kinship norms influence how the immigrants interpret and adapt to
adversities, inequalities, and new identities in their host society (p.7).

Although Hein (2006) conceptualizes and uses ―ethnic-origins‖ to examine racial and
ethnic adaptation drawing parallels between the transplantation of ethnic boundaries (the
cohesiveness of the group) and the propensity for mobilization, he also stressed that ―all forms of
immigrant adaptation involve an interaction between immigrant conditioning in their country of
origin and conditions in the host society‖ (p.25). It is from this acknowledgement – that an
immigrant‘s culture and tradition play a continuing and vital role in various aspects of immigrant
adaptation – that I base my hypothesis of clan allegiance influencing the spatial distribution of
Somalis in America. The major contribution of the theory of ethnic-origins is that it shows that a
group‘s origin and cultural past can be used to explain the basis of many of their actions in the
present. The major drawback of the theory, however, is that it is more suitably applied to first
generation immigrants.

17

Behavioral Approach to Migration
In examining Somali motivations for migration, this research follows a behaviorally
oriented approach to migration – this is an approach that attempts to replace the concept of the
economic man9 and its assumptions of profit maximization with more realistic motivations for
migration. According Cadwallader (1992), any attempt to analyze migration using the behavioral
approach should incorporate the human capital theory; therefore, this research will be grounded
by the human capital theory of migration where migration is seen as an investment in the future.
The human capital perspective draws attention to how the decision to migrate is made at the
microlevel – individuals consider the cost and benefits of migration and migrate if they believe
the benefits outweigh the costs (Gurak and Kritz 2000). Sjaastad (1962) was among the first to
recognize the human agency and the temporal element when trying to account for differential
earnings in light of net migration and theories of labor market equilibrium; migration should be
treated as a time dependent investment where the benefits of a move are not immediately
noticed. The investment is seen as increasing the human potential; therefore, it is an investment
in human capital because the investment increases the productivity of human resources (Clark
1986). Proponents of this behavioral approach, ―have been able to demonstrate that a wide range
of social phenomena can be attributed to individual decisions involving the postponement of
present gains for the expectation of future gains‖ (Cadwallader 1992, 115). Cadwallader also
recognized that the human capital model not only incorporates the financial aspects of migration,
but also the nonfinancial aspects such as expected gains in amenities and services. In this light,
this research incorporates a wide spectrum of independent variables that reflect the spirit of the

9

Economic man is defined as an imaginary ‗perfectly rational person‘ who, by always thinking
marginally, maximizes his or her economic welfare and achieves consumer equilibrium
(Businessdictionary.com 2010).
18

behavioral approach‘s assertion of the value of nonfinancial aspects as possible determinants of
migration.

Network Approach to Migration
Another aspect that should be considered while studying the motivation for Somali
migration in the United States is the network approach. This approach places emphasis on
migration being imbedded in a series of ethnic, political, familial, and communal relationships
(Gold 2005). The network approach is highly applicable to the Somali in light of the prevalence
of clan-based allegiances that permeates Somali society – both at home and abroad.

An

investigation of networks allows for the understanding of the migrant‘s immersion in multiple
levels of collective connections (p.277). Networks of relations provide for potential and actual
resources, like access to jobs, which suggests that they are a form of social capital (Raghuram,
Henry and Bornat 2010). Social capital is defined as ―a person‘s or group‘s sympathy towards
another person or group that may produce a potential benefit, advantage, and preferential
treatment for another person or group of persons beyond that expected in an exchange
relationship‖ (Robison, Schmid and Siles 2002, 6). Social capital denotes a web of loyalties and
obligations that develop between people leaving a sense of commitment that induces the
extension favors and expectation of preferential treatment (Gold 2005). This form of capital,
therefore, enables individuals to command scarce resources by virtue of their membership in
social networks or broader social structures (Portes 1995).
The preceding concepts are used to guide this research by informing of what to look for
in the analysis of the subject group. For example, the ethnic-origins hypothesis posits that
immigrant groups will apply traditions from their country of origin as a method of adaptation in

19

their new environment. In Somalia, the nomads relied on the ancient tradition of sahan where
they would send out scouts on reconnaissance to find suitable pastures and water for livestock.
When the scouts reported back from their exploratory expedition with news of desirable pastures
and other relevant information for survival, camp would be broken and the nomads would
migrate to that area (Abdullahi 2001). The tradition of sahan has been transplanted from
Somalia and put into practice in the United States, but modified for a more modern environment.
According to Tilove (2002), even though the Somali reconnaissance of Lewiston took place via
Internet, telephone, and air travel, it all began with the Somali concept of sahan. Powell (2002),
in an article for the Washington Post describing the events that took place in Lewiston, wrote
that the Somali sent out seven men in the four cardinal directions from Atlanta, GA. Once they
found Lewiston, they researched its crime statistics, unemployment rate, and housing. The men
reported back to the elders in Atlanta that Lewiston was ―a dream place‖ and by February of
2001, the first group of Somalis was on a Greyhound bus headed for Lewiston (Powell 2002).
This event serves as a prime example of immigrants applying traditional ways in a new
environment as posited by the ethnic-origins hypothesis.

Research Questions
The overall goal of this research is to gain insight into the Somali population here in the
United States. This will be achieved by examining the following research questions; (1) is
Somali migration(s) related to clan more so than other socio-demographic variables? (2) Are
clans establishing areas of dominance that could affect the future flow of migration?

For

example, would members of the Somali clans who still adhere to clan traditions avoid certain

20

cities10 if they know that members of a rival clan make up the majority of the Somali population
in that city? Some of the specific questions that this research will address include; (1) are the
cities where the Somali are initially resettled diversified with respect to clan-families? (2) Have
the Somali dispersed from their original sites of resettlement? (3) What socio-demographic
factors are associated with the distribution of Somalis? (4) What socio-demographic factors are
associated with Somali migration within the United States? (5) Are there cities associated with
specific clans and what are the factors related to that association. It is important that we
understand the motivations for Somali migration because, as mentioned earlier, the Somali are a
growing segment of our population and given the continuing instability in Somalia, the Somali
population in the United States will continue to grow via new infusions from Somalia well into
the foreseeable future.

Data and Methods
To answer the questions mentioned above, this research will draw upon both quantitative
and qualitative data and methods. The quantitative data was acquired from a variety of public
databases and analyzed using a variety of statistical tools. The qualitative aspect is delivered via
participant observation, surveys and unstructured interviews designed to collect information
from both open and close ended questions. The predominant reason for mixing the qualitative
with the quantitative – as noted by Bryman (2006) – is for comprehensibility of research findings

10

According to the U.S. Census, a city is defined as a type of incorporated place in the 49 states
and the District of Columbia. An incorporated place is a type of governmental unit incorporated
under state law having legally prescribed limits, powers, and functions
(www.census.gov/geo/www/geo_defn.html#Place).
21

with one method enhancing the results of the other method. To answer the research questions, I
arranged the methodology in a logical progression consisting of five steps.

In the first step, I will use Simpson‘s Index of Diversity and data from the ORR‘s
Worldwide Refugee Admissions Processing System (WRAPS) to determine if the cities where
the Somali are initially resettled diversified with respect to clan-families. I hypothesized that
these resettlement cities were originally highly diversified with respect to the settlement of
Somalis from different clan-families. The logic behind this step is to show the degree of
diversity which will lead to hypotheses regarding motives for the Somali secondary migrations
given the history of conflict between the clans in Somalia. In the second step, I will use data
from the WRAPS in conjunction with data prepared by the U.S. Census Bureau from the 2000
Census to calculate a relative clustering index to determine if Somalis are dispersing from their
original settlement sites. The relative clustering index is a measure of clustering; however,
knowledge about Somali dispersion will be gained by analyzing the clustering. I hypothesize
that Somalis disperse from their initial resettlement communities therefore becoming more
diffused throughout the United States. Assertions have been made about the changing geography
of the Somali, but to this researcher‘s knowledge, the assertion has never been empirically tested.
I posit that when some Somali are resettled, they are resettled in areas populated by rival groups;
the inherent tension facilitates the revival of traditional behaviors whereby the reliance on clan
for safety and security is necessitated. With clan allegiance rekindled, the group migrates to
communities that are more favorable to their growth and prosperity. Steps three and four involve
the determination of social, economic, and demographic factors associated with the distribution
of Somalis. In the third step I will use a multitude of social, economic, and demographic

22

variables from several government databases and subject them to a principal components
regression (PCR). The goal of this step is to determine the order of factors that may have the
most influence on the residential distribution of Somalis across the American landscape. The list
of variables will not be all-inclusive, but it will contain over thirty variables based on eleven
aspects I believe to be important in the residential distribution of the Somali. The aspects will
range from crime to public housing to employment. This process will be repeated in the fourth
step in order to determine the factors that may have the most influence in determining Somali
migration destinations.

The fifth step in this research deviates from the first four in that it will be based on the
collection of primary data from the Somalis migrants themselves versus the use of secondary
data. In this step, I will rely on a collection of qualitative data from participant observation,
unstructured interviews, and surveys designed to elicit information not contained in national
databases.

Information pertaining to clan-family membership from Somalis admitted for

resettlement is collected as part of the initial refugee admissions process – clan membership is
used to determine whether the applicant has a profound fear of persecution (Horst 2006).
Because of this aspect of the admissions process, there is a count of the Somali admitted to the
United States by clan membership. From this admission information, an assessment of the initial
distribution of the Somali by clan-family can be made. Outside of the information on clanfamily membership collected on entry, there are no other procedures (census or otherwise) which
request clan-family affiliation information from the Somali which means there is no way to
assess the migration patterns of Somali by clan-family membership in the United States via an
established database. Because of this lack of information, surveys, interviews, and participant

23

observation will be used to obtain knowledge about clan-family membership of the Somali
migrants. Specifically, these instruments will be used to determine if there are cities in the
United States associated with specific clan-families. These instruments will also be used to
identify the factors related to that association. The first survey will be designed for and sent to
non-governmental organizations (NGOs) that provide refugee services. This survey will be sent
to NGOs all across the United States to solicit information about their Somali secondary
migrants. The second survey will be localized. This survey will be specifically designed for
Somali secondary migrants and will be used in combination with interviews and participant
observation in the study area. The data and methods will be discussed in detail in chapter 5.

The Study Area
Columbus, Ohio was chosen as the site for my field research primarily because of the size
of its Somali population. For the 2000 census, there were 2,839 Somali living in the City of
Columbus (U.S. Census Bureau 2000). The current estimates of the Somali population living in
Columbus, however, vary from 30,000 – 50,000 making it the second largest concentration of
Somalis in the country behind Minneapolis, Minnesota (Ohio Homeland Security 2010).
Another reason for choosing Columbus as a study area was because it was recently the
destination point for a group of 300-plus Somali secondary migrants (Ohio Refugee Services
2006). Both of the aforementioned reasons made Columbus the ideal study area because of the
inherent effect on widening the potential pool of subjects.

24

Significance of the Study
This research differs primarily in two ways from other studies of immigrant locational
determinants. The first concerns the group to be analyzed and the second concerns the method
of analysis. First of all, to this researcher‘s knowledge, no one has quantitatively analyzed the
locational determinants of the Somali population in the United States.

Secondly, I use a

Principal Components method of analysis instead of the linear regression in identifying
locational determinants. This method was chosen because it allows for the retention of more
variables without succumbing to the bias introduced via tolerance issues when analyzing
sociological data (Jolliffe 2002).

The importance of this research is that it can serve as an initiating point for further
research in other countries included in the Somali diaspora. By focusing on the spatial aspect of
the Somali population, comparisons can be made which can be included in the overall
framework for studying the Somalis in the global diaspora. This research also has the potential
to effect policy with respect to the initial placement of Somali refugees. If it is discovered that
Somali secondary migrations are primarily due to the desire to reconstitute communities based
on kinship and clan rather than country of origin bonds - then the impetus to rethink the
resettlement procedures with respect to Somali refugees might be warranted. Media reports
associated with the case studies incorporated into this introduction indicate that when the Somali
migrate, a substantial amount of resources are needed to help them settle into their new
environment – by re-thinking resettlement procedures, the loss of valuable resources might be
avoided. The overall significance of this research is that it will add to the small body of
literature that currently exists on the Somalis living in the United States.

25

Summary of Remaining Chapters
While the preceding passages briefly highlight some of the themes and mechanics of this
research, a more in-depth review is needed to fully understand the underlying logic. In the
second chapter, I focus on the Somali; in line with the theory of ethnic origins, a thorough
understanding of Somali history and conflict help to form the basis for understanding their
motives and actions in the present. In this chapter, I discuss aspects of Somali culture, the post1960 events which facilitated the flow of refugees from Somalia, and Somali resettlement in
America. In the third chapter, I explain in detail the concept of refugees and refugee resettlement
– an understanding of the resettlement process is necessary for an understanding of the numbers,
distribution, and general behavior of these new Americans. The fourth chapter is reserved for
the literature review. In the fourth chapter, I examine the research that has been done on the
Somali and refugees in general; I make the distinction between the two because of the themes in
the literature – research on the Somali neglect locational determinants and research on locational
determinants neglect the Somali. The remaining chapters – five through seven – are devoted to
research methodology, research findings, and finally, the conclusions drawn from this research.

26

CHAPTER 2: THE SOMALI

This research examines the distribution of the Somali in the United States by referencing
prominent feature of their society in an attempt to uncover its influence on their migratory
practices. While there is little acknowledgement that clan allegiance is reconstituted in the
Somali global diaspora, there is overwhelming evidence of its past and present effect on the wartorn region of Somalia. In this chapter, I introduce Somalia and Somalia‘s history to put the
influence that clan has on the Somali populace into perspective.

Somalia
Geography and Climate of Somalia
Somalia is a country on the far eastern coast of the African Continent in the geographic
formation aptly known as the ‗Horn of Africa‘ because the physical change between the land and
the sea (Figure 2.1) is reminiscent of the horn of a rhinoceros (Alexander 2001). Somalia‘s land
area is a little less than 638,000 square kilometers – which is a little smaller than Texas, but a
little larger than France – and is bordered by Djibouti and the Gulf of Aden in the north, Ethiopia
in the west, Kenya in the south, and the Indian Ocean in the east. Somalia is mostly flat plains
and plateaus with mountainous areas to the north. The weather is mainly hot and dry year-round
with exception to the northern areas of higher elevation. There are two wet seasons, which bring
erratic rainfall to Somalia from April to June and from October to November. The environment
is arid to semi-arid which facilitates the nomadic pastoralism that more than half the population
practices (Putnam and Noor 1993).

27

Figure 2.1: Geographic location of Somalia. For interpretation of the references to color in
this and all other figures, the reader is referred to the electronic version of this dissertation.

28

According to Lewis (1982), because of the extreme environment, nomadism became the
prevailing economic response and mode of livelihood to adjust for the scant resources of the
unenviable harsh environment. Climate is the motivating factor influencing much of Somali life;
for the largely nomadic population, the timing and amount of rainfall serve as crucial elements in
determining the adequacy of vegetation for grazing and thus the relative prospects for prosperity
(Alexander 2001). The four seasons recognized by the Somali include: two rainy seasons (gu
and day); and two dry seasons (jiilaal and hagaa). The gu rains occur briefly turning the vast
desert regions and the plateau into lands suitable for grazing with the growth of lush vegetation.
This season is followed by the hagaa season, which is typified by drought conditions between
July and September, which is thus followed by the day rains. The time between December and
March is known as the jiilaal season and is classified as the harshest season for pastoralists and
their herds (Alexander 2001). Somalia receives less than 19.7 inches of rain on average annually
– the north and northeastern areas receive as little as 2 to 5.9 inches, but in the southern region,
the average annual rainfall ranges anywhere from 12.9 to 19.7 inches. Mean temperatures in
Somalia range from 68 to 86 degrees Fahrenheit; the greatest temperature extremes, however,
occur in Northern Somalia with temperatures ranging from below freezing in the highlands
during jiilaal to 113 degrees Fahrenheit during hagaa.

The Somali People
A 2007 estimate placed Somalia‘s population count at 9.1 million – this was based on
extrapolations from a 1975 census taken by the Somali Government. Most calculations of the
people in Somalia are estimates based on this year because the enumeration process is hindered
by the large number of nomads and by refugee movements in response to famine and clan

29

warfare (The World Factbook 2008).

While the Somali are typically described as being

homogenous11, there are a number of segmentations that differentiate the people both ethnically
and culturally. Figure 2.2 represents the basic structure of that segmentation found in Somali
society.

The Somali come from a clan-based society – kinship groups, or lineages, form the basic
building blocks of Somali society and the basis for Somali identity (Lewis 2004). The clanfamily represents the upper limits of clanship; however, due to the size of the grouping, it is
impossible for the clan-family to serve as a political unit. The clan, as a much smaller grouping,
is more manageable and thus marks the upper limits of the political unit. At the base of the
lineage system is the dia-paying group – the basic political and jural unit where ―contract and
clanship meet.‖ (Lewis 1982, 6). The preceding statement is in reference to the notion that the
two concepts are fundamental principles underlying Somali political units (Lewis 1982). The
largest segmentation of the Somali people is between the Samale and the Sab. Brons (2001)
classifies this segmentation as the basic divide between the nomadic and the settled Somalis.
According to the Brons, the division between the Samale and the Sab has a discriminatory
connotation that still affects Somali society and politics by influencing power relations and
conflicts because the Samale majority promoted the nomadic lifestyle as noble and created a
national identity that marginalized the sedentary Somali – i.e. the Sab. The Samale make up the
largest proportion of the population – they are primarily pastoral nomads and are distributed
throughout the land (Lewis 2004).
11

The Somali are considered homogenous; however, there are numerous minority groups also
present in Somalia to include the Bantu, the Bajuni, the Barawans, and the Benadiri. The
country itself is 85% ethnic Somali and 15% Bantu and other non-Somali (The World Factbook
2008).
30

Figure 2.2: Basic structure of Somali segmentary lineage.
ClanFamily

Clan

Clan

Clan

Sub-clan

Sub-clan

Primary Lineage

Primary
Lineage

Dia-paying
Group

Dia-paying
Group

The Samale are divided up into four large clan-families: the Dir, the Darood, the Isaaq,
and the Hawiye. The Dir are found in the extreme north and in sections of the south. The
Darood can be found in the eastern parts of the north and the central regions of the country. The
Isaaq predominantly occupy the northern regions, and the Hawiye are predominately found in
parts of the central region with extensions to the south (Abdullahi 2001). The Sab, on the other
hand, is composed of a related group of peoples who live in southwestern Somalia in the riverine
area. The Sab is divided up into two main clan-families – the Digil and the Mirifle – who are
sometimes collectively known as the Rahanweyn (Abdullahi 2001). The Sab, many of whom
have more prominent African features, are to a certain extent considered inferior by the Samale
who claim to be the original Somali and, with that, claim a right to cultural and political
domination (Brons 2001). In addition to the main division – Samale and Sab – based on

31

production/settlement patterns12 and the division into clan-families, the Somali lineage system is
continuously segmented into dozens of clans, and hundreds of sub-clans. Figure 2.3 is a map
representing the distribution of some of the clan-families13 as of year 2006 – this distribution,
however, is subject to change as a result of the on-going conflict. A major reason for the
incorporation of this map is to support the statements made about the territorial nature of the
Somali in their homeland as mentioned earlier in this dissertation. One special point of interest
about the distribution of the clan-families is their occupation of lands crossing regional and
international boundaries. Territorial borders have been imposed on the Somali, but they have
little meaning especially to the nomadic clans in Somalia (Horst 2006).

Appendix Tables A.1 and A.2 contain a partial listing of the numerous clans and subclans found in the Somali region. Identifying clans constitutes a complex task, and it would be
almost impossible to construct a complete table of the clan groupings because they are akin to
living organisms subject to constant division and development (Yoshimura 2009). Thus, like the
general areas occupied by clans, which is subject to change, the clan hierarchy is also subject to
change with new divisions and segments forming regularly. Therefore, the tables presented may
not represent a full account of all the groups present in Somalia. The primary purpose in
including such a table is to relay the complex diversity and segmentation that is found within the
Somali society. However, despite the society being organized according to the principles of
segmentary lineage, there is the common link of Islam between virtually all groups in Somalia.
The Islamic faith serves as a horizontal identity that bisects clan alignment (Le Sage 2001).
12

The Samale are predominantly nomadic pastoralist while the Sab are typically settled farmers.
In the case of the Isaaq, there is debate as to whether they constitute a clan-family of their own
because the Southern Somalis and the Majerteen clan of the Darood claim the Isaaq are a part of
the Dir (Yoshimura 2009).
13

32

Figure 2.3: Map of the distribution of Somali clan-families in Somalia.

33

The Arabs brought Islam to Africa in the seventh century and the majority of the Somali
population had been converted by the year 1100. By the mid-fifteenth century, the Somali had
joined with the Arabs in fighting the Islamic holy wars against Ethiopian Christians (Advameg
2011). Thus, in pre-colonial times, the Somali community recognized two main authorities –
clan elders and religious leaders (Le Sage 2001). The affairs of the leaders typically overlapped
to the extent that Islam (which nearly all Somali practice) was essentially assimilated into clan
culture in a symbiotic relationship that has lasted throughout the colonial and well into the postcolonial period (p.472).

Despite the relative religious and ethnic homogeneity14 reported amongst the Somali,
Somalia has a history of conflict between its clan-families, clans, and smaller sub-clans (Schaid
and Grossman 2007). According to Lewis (1982), the first mention of Somali clans appeared in
the mid-thirteenth century when the Arab geographer Ibn Sa‘iid mentioned the Hawiye near
Merca in southern Somalia. The first mention of clan conflict, however, did not occur until the
sixteenth century when Arabian historian Shihaab ad-Diin wrote in his Futuuh al-Habasha
(‗History of the Conquest of Abyssinia‘) about the frequency of quarrels and struggles between
the Somali lineages to which Lewis notes ―took a similar course to that which they follow today‖
(p.16). Logic suggests, and history confirms, that the conflict and division between the clans and
sub-clans would become more defined as an after effect of the colonization in light of the
differential treatment and skills the various clans and sub-clans received while subordinate to
three opposing colonial powers.

14

According to Abudullahi (2001), the Somali are 100 percent Sunni Muslims. In addition, it is
reported that over 85 percent of the Somali population are ethnic Somalis.
34

The Colonization of Somalia
By the middle of the 19th Century, Somalia was in the process of being partitioned up
primarily by three colonial powers – Britain, France, and Italy (Figure 2.4). The occupation
officially began with Britain. In an effort to secure the northern coast from their European
competitors, after the opening of the Suez Canal, Britain signed treaties of protection with many
of the northern Somali clans between 1884 and 1888 (Abdullahi 2001). Britain‘s primary
interest was in Aden – which it took possession of in 1839. Britain‘s establishment of a garrison
at Aden served as a station for ships sailing between India and Europe (Besteman 1999). With
the establishment of the garrison, there was a strategic importance to Britain‘s interest in Somalia
– however, some believe ―the Somali region never played more than a secondary role for the
powers involved in colonial occupation‖ (Brons 2001, 130). Some twenty-two years after the
British secured Aden, France followed suit by securing the Port of Obock and extending its
protectorate over a small block of land inhabited by the Somalis. The Italians soon followed the
French as they secured the Port of Asab in 1869. Many point to this occupation by imperial
forces as one of the root causes of the conflict that tore the country apart. When colonized
countries gain independence, the burdens of the inhabitants are typically far from over due to the
lingering effects the colonizers have left. The colonization process not only creates changes in
the colonized culture, but also changes are commonly made to spatial boundaries of the
colonized regions and the political philosophies governing the people of that land. Somalia
represents a special case in that the country was divided up not by a single colonizing power, but
by three major colonizing powers.

35

Figure 2.4: Colonial partitioning of Somali territory mid-19th century.

36

In addition to the major colonial powers, there were also local powers15 involved in the
disection of Somalia. One of those powers included Ethiopia which, at that time, was under the
rule of Emperor Menelik II. Ethiopia constituted a formidable opponent because Emperor
Menelik and his forces were able to thwart any incursion of colonial powers into Ethiopian
territory.

Ethiopia also competed with the colonial powers for partitions of Somali land

(Alexander 2001).

The Ogaden region (land acquisitioned by Ethiopia during Somalia‘s

occupation) serves as a prime example of the lingering effects of colonization because Somalia
would later be involved in a war with Ethiopia16. The foundation of the war was based on the
attempt to reclaim this land (the Ogaden), which was originally a part of Somalia prior to
colonization.

The Aftermath of Colonization
British Somaliland was awarded its independence on 26 June 1960 and united with
Italian Somaliland to establish the Somali Republic on 1 July 1960 (Advameg 2011). Even
though the British north and the Italian south were again united, the country was far from being
unified. Italy and Britain had left the Somalis with two separate legal and administrative systems
that conducted business in different languages using different procedures (Alexander 2001). The
two regions held divergent views on most aspects. The opposing views were bolstered by the
assertions of the southern Somalis who insisted that their education and preparation under Italian
rule made them best suited to run the country. Thus, communication between the North and the
South was virtually nonexistent. The lack of consensus between the two regions necessitated
intervention by the United Nations, which led to the creation of the Consultative Commission for
15
16

The other countries bidding for Somali lands included Egypt and Ethiopia.
The Ogaden War fought between 1977 and 1978.
37

Integration in order to bring the two halves together (Alexander 2001). The problem with
political cohesion in Somalia stems from representation. As mentioned previously, the basic
building blocks of the society was clan-based, and this became evident in the political arena after
independence. The creation and splintering of clan-based political factions along with the
formation and dissolution of clan allegiances increased exponentially. Table 2.1 shows the
changes in factions holding political seats between the 1964 and 1969. In 1969, the increase to
60 political parties scrambling for the 123 parliamentary seats marred the elections with
corruption, violence, chaos, and corruption. Electoral violence led to the eventual assassination
of President Abdirashid Shermarke bringing the country to the edge of anarchy (Adam 2008).

Post-Independence: The Siyaad Barre Years
During the turmoil in 1969, General Mohamed Siyaad Barre staged a bloodless coup after
President Abdirashid Shermarke was assassinated.

Shortly after seizing power, the new

president stood on a platform of ‗Scientific Socialism17‘ and stated a goal of uniting the Somali
nation and eradicating ancient clan divisions (Lewis 1994).

According to Brons (2001),

clannism was officially banned – President Barre‘s administration attempted to sanitize social
life of its clan foundations, and any attempts to reconstitute clan or any form of clan allegiance
was punishable by death. In fact, during President Barre‘s reign, ―imprisonment, extra-judicial
killings, torture, and other widespread violations of human rights accompanied political control‖
(p.178). While trying to suppress the bonds of clan and promote nationalism, however, President
Barre covertly relied on clan loyalties to build his own inner circle of power (Lewis 1994).

17

The Somali practice of socialism became official on the first anniversary of the coup led by
General Barre.
38

Table 2.1: Somali political factions 1964 - 1969.
30 March 1964 National Assembly Election
Party
Number of Seats (123)
Somali Youth League (SYL)
69
Somali National Congress (SNC)

22

Somali Democratic Union (SDU)

15

Somali Independent Constitutional Party (HDMS)

9

United Somali Party (USP)

1

Liberal Somali Youth Party (PLGS)

1

Somali National League (SNL)

1

Somali African National Union (SANU)

1

Others

4

26 March 1969 National Assembly Election
Party
Number of Seats (123)
Somali Youth League (SYL)
73
Somali National Congress (SNC)

11

Somali Independent Constitutional Party (HDMS)

8

Somali African National Union (SANU)

6

Liberal Somali Youth Party (PLGS)

3

Somali Democratic Union (SDU)

2

Popular Movement for Democratic Action (PMDA)

2

Somali Socialist Party (PSS)

2

Somali People's Movement Party (PMPS)

2

Revolutionary Socialist Workers' Party (PRSO)

1

Somali National Solidarity Party (PSNS)

1

United Somali Party (USP)

-

Others

12

Source: African Elections Data @ http://africanelections.tripod.com/so.html

39

Barre‘s ring of power revolved around three relative clans from the Darood clan-family:
Marehan – which was the president‘s own clan, Ogaden – which was the president‘s mother‘s
clan, and the Dulbahante – which was the president‘s son-in-law‘s clan. These three clans –
collectively known as the M.O.D. – exercised special power because of their relation to President
Barre (Lewis 1994).

Somali Nationalism
According to Besteman (1999), one of President Barre‘s greatest aspirations was to create
a united Somali and to include the Ogaden region in that unification. The Ogaden was under
Ethiopian control – but more importantly, the Ogaden was the region of President Barre‘s
mother. Pan-Somalism was the name given to the concept underlying the hope of the unification
of all lands populated by Somalis, and it was one of the most important political issues during
that time. Preoccupation with the concept shaped the character of the country‘s newly formed
institutions, extended the popularity of President Barre, led to the build-up of the Somali
military, and ultimately led to the war with Ethiopia (Alexander 2001). In 1977, the Somali
invaded the Ogaden region to reclaim the land which was once theirs prior to colonization, but
was forced to retreat when the Soviets switched their backing to the new Marxist Ethiopian
government with military support against the Somali incursion (Besteman 1999). The war over
the Ogaden region was immensely popular in Somalia at first, and President Barre‘s public
standing had never been higher than at that time, but the terrible defeat ―quickly led to
widespread public demoralization and to an upsurge of ‗tribalism‘ as different groups sought
scapegoats to explain the debacle‖ (Lewis 1989, 575). Many believe the loss of the Ogaden War
signified the beginning of President Barre‘s decline in popularity, his loss of ideological vision,

40

his acceleration of the manipulation of clan politics, and his rise in tactics of state repression
(Besteman 1999). Somalia faced a humiliating loss to Ethiopia, which created an influx of
Ogaden refugees into Somali controlled territory exacerbating clan tensions. Awareness of the
unequal distribution of state resources based on clan favoritism was becoming a problem. This
was especially true in northern Somalia where the Isaaq clan-family was increasingly becoming
excluded from state resources in favor of the Ogaden refugees who were backed by the M.O.D.
(Besteman 1999).

The Somali Conflict
Escalating conflicts over resources eventually progressed to all out war between clans
and factions within clans – Barre‘s regime fell in 1991. Mohamud and Kusow (2006) believe
that the people of Somali anticipated better leadership and government after the fall of the Barre
regime but instead they got warlordism18. According to Lewis (1989), in their desperate fight for
survival, the Barre family and clansmen sought to exploit the full segmentary lineage rivalry
within Somalia. Conversely, the continued clan conflict and violence, in conjunction with
drought and famine, produced and continues to produce hundreds of thousands of refugees –
many of whom eventually became resettled throughout the world in a global diaspora. Table 2.2
represents the top five refugee sending countries between 1997 and 2009. Out of the thirteen
years depicted, the Somalis have occupied a place in the top five refugee sending countries for
nine of those years. For three of the years that they did not rank in the top five, the Somalis still
ranked in the top eight.

18

Warlordism is a style of leadership where influential individuals appeal to the sentiments of
their clan or sub-clan members in order to advance their own political and economic interests.
41

Table 2.2: Top five refugee group admissions to the United States 1997 - 2009.
Rank:
1
2
3
4
5
Rank:
1
2
3
4
5
Rank:
1
2
3
4
5
Rank:
1
2
3
4
5
Rank:
1
2
3
4
5

COUNTRY
Russia
Bos.-Herzegov.
Vietnam
Somalia
Cuba
COUNTRY
Bos.-Herzegov.
Ukraine
Somalia
Iran
Sudan
COUNTRY
Ukraine
Liberia
Iran
Sudan
Somalia
COUNTRY
Somalia
Russia
Cuba
Vietnam
Iran
COUNTRY
Iraq
Burma
Bhutan
Iran
Cuba

FY1997
27,072
21,357
6,660
4,974
2,911
FY2000
19,033
7,334
6,026
5,145
3,833
FY2003
5,065
2,957
2,471
2,139
1,994
FY2006
10,357
6,003
3,143
3,039
2,792
FY2009
18,709
18,275
13,317
5,374
4,800

COUNTRY
Bos.-Herzegov.
Russia
Vietnam
Somalia
Iran
COUNTRY
Bos.-Herzegov.
Ukraine
Iran
Sudan
Somalia
COUNTRY
Somalia
Liberia
Laos
Sudan
Ukraine
COUNTRY
Burma
Somalia
Iran
Burundi
Cuba

42

FY1998
30,906
23,349
10,288
2,951
1,699
FY2001
14,593
7,172
6,590
5,959
4,951
FY2004
13,331
7,140
6,005
3,500
3,482
FY2007
13,896
6,969
5,481
4,545
2,922

COUNTRY
Bos.-Herzegov.
Serb. and Monte.
Vietnam
Ukraine
Russia
COUNTRY
Ukraine
Bos.-Herzegov.
Vietnam
Russia
Cuba
COUNTRY
Somalia
Laos
Cuba
Russia
Liberia
COUNTRY
Iraq
Burma
Thailand
Iran
Bhutan

FY1999
22,699
14,280
9,622
8,649
4,386
FY2002
5,216
3,461
2,988
2,105
1,919
FY2005
10,405
8,517
6,360
5,982
4,289
FY2008
13,755
12,852
5,279
5,257
5,244

The influence of the many push factors (primarily war, but also famine and drought) from
Somalia is also aptly captured by the graph in Figure 2.5. This graph indicates (by looking at the
peaks which correspond mainly to the escalation of fighting in the region) that the Somali
comprised the majority of African refugee admits the first time between 1994 and 1997.

The

second time they occupied this position was between 2005 and 2006. For the 1990 to 2009 time
frame, 41.3 percent of the African refugees admitted to the United States were from Somalia.

Figure 2.5: Percentage of refugees from the African region who are Somali.
PERCENTAGE OF AFRICAN REFUGEES FROM
SOMALIA 1990 - 2009
100.0%
90.0%
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
SOMALIAN

Barriers to Unification
According to Lewis (1963), the Somali never constituted a single autonomous political
unit – before the partition of their lands by Britain, France, Italy, and Ethiopia, the country was
divided into numerous hostile clans. The Somalis did have a common code - they had ―a

43

sentiment of Somali-ness, accompanied by a virtually uniform national Somali culture, and
reinforced by the strong adherence of all Somali to Islam‖ (p.147). So the Somali essentially
constituted a nation19, but ―political nationalism was largely absent because of the divisive forces
within the nation‖ (p.147). The divisive forces, more often than not, is rooted in the prevalence
of clan and the associated obligations.

Mohamud and Kusow (2006) identified several factors that laid the foundation for the
current political crisis and dissolution of the country. The first is the ongoing conflict with
Ethiopia that began in the sixteenth century with the Islamic and Christian wars. The second
factor hindering the Somali political process and sustaining the inability of Somali politicians to
resolve the ongoing national crisis is the divisive politics of clan ideology (p.16). According to
Osman (2006):
Over the years the control of the state has been viewed as zero-sum game where the
winners take all. Thus, it has become a competition to the bitter end for certain Somali
clans where on the one hand they fight for the control of the state and on the other hand
prevent others from its control. The method used to enforce the unequal distribution of
the state benefits have always been clannism, tribalism and regionalism. (p.74)
While the persistence of clan has consistently been a barrier to national unity, in most cases, the
barrier would have emerged before the consolidation of a region into a single political unit. In
the case of the Somalis, it came afterwards – as ―soon as the political machine of postindependence Somalia started turning, the achievements of nationalist and independence
movements gave way to a more divisive clan politics‖ (p.18).

19

The definition of nation for our purpose is defined as a culturally distinctive group of people
occupying a particular region and bound together by a sense of unity arising from some shared
ethnicity, beliefs, and customs (Getis, et al. 2011).
44

Somali 4.5 System
One of the lasting effects of the Barre regime was that it showed the Somalis that the
state and its resources belonged to those in power and the clans of those in power. Those who
were kept at bay by those in power, realizing that they have no representation and therefore no
rewards associated with the state, intuitively learned to equate the concept of state with clan or
clans that occupied the political seats of that state. The convergence of the concepts of clan,
representation, and politics is an issue that has yet to reach equilibrium in Somalia.

Since the fall of the state in 1991, the founding of the Transitional Federal Government
(TFG) – some 15 reconciliation attempts and 13 years later – marked the acknowledgement of a
will to accommodate the complex historical, social, and political concerns of the Somali. The
TFG‘s charter recognized the prevalence of clan in Somali politics and sought to find some kind
of equilibrium with the 4.5 formula which emerged during the Reconciliation Conference in
Djibouti in 2000. According to Hassan and Barnes (2007), the TFG institutions are reportedly
based on the so-called ‗4.5 formula‘ which was designed to balance and share representation and
power in Somalia between four main clan families (Dir, Darood, Hawiye and Rahanweyn), as
well as five minority constituencies. Hassan and Barnes assert that the TFG‘s constituent parts
appear to reflect the ―4.5 formula‖ superficially, but later admit that the perception of many in
southern Somalia is that the TFG is primarily a Darood institution. According to Eno and Eno
(2009), the 4.5 system can be interpreted as the separation of the Somali people into four clans
that are considered ―pure‖ Somali and therefore equal (the ―4‖ from the formula). This is in
opposition to an amalgamation of various minority clans and groups that are unequal (the ―0.5‖
from the formula) and therefore ―impure‖ or less-Somali. Thus, from the start, the 4.5 formula

45

represented a system of ―absolute discrimination and severe ethnic marginalization‖ (Eno and
Eno 2009, 138).

While there are those who would applaud the 4.5 formula and bill it as a big step in
equalizing the power distribution amongst the various clans at the state-level, others would
denounce it as legalized discrimination and a show of blatant ethnocentrism. The groups that
constitute the ‗.5‘ have little representation – and therefore little power.

The groups that

constitute the ‗4‘, who have equal representation and the associated power, would push for more
– or even all – of the power. This notion is aptly captured in Brian J. Hess‘s (2010) analogy
linking the Somali politics and power scene to the Somali parables of predation:
The lion, the jackal, the wolf and the hyena agreed to hunt together, and to split their
kills. After cooperatively killing a camel, the lion asked ‗Who divides the meat?‘ The
wolf volunteered to divide the meat, as he could count. He promptly cut four pieces, each
of equal size, and placed the pieces in front of the hunters. Angry, the lion swiped the
wolf across his eyes with his massive paws and claws causing the wolf to cower in
bloodied pain. ‗Is this any way to count?‘ roared the lion. The jackal intervened. ‗The
wolf does not know how to count. I will divide the meat‘. The jackal then cut three small
portions for the wolf, the hyena and himself. He placed the largest, best piece in front of
the lion, who promptly collected his share and departed. The wolf, the hyena and the
jackal were left with their relative scraps. ‗Why did you give the lion such a large piece?
Why is he entitled?‘ asked the hyena. ‗I learned from the wolf‘, replied the jackal. ‗You
learned from the wolf?‘ cried the hyena. ‗How can anyone learn from the wolf? He is
stupid‘. Still nursing his injured eyes, the wolf ended the debate: ‗The jackal is right. He
knows how to count. Before, when my eyes were open, I did not see. Now, though my
eyes are wounded, I see clearly‘. (p. 254)
Hess‘s rendition of the Somali parable includes four main characters symbolizing the four groups
currently sided against the TFG who are not content with their share of the power and desire
more – if not all: (1) al-Shabaab, primarily drawn from the Hawiye clan; (2) Hizbul Islam, with
militiamen drawn primarily from the Darood; (3) Raas Kaambooni, with militiamen drawn
primarily from the Mohamed Subeer subclan of Darood; and (4) Anoole/al-Furqaan with

46

militiamen drawn primarily from the Majeerteen, Warsengeli, and Dhulbahante sub-subclans of
Darood (Hesse 2010, 254).

The tail end of the graph in Figure 2.5 displayed earlier shows a sharp climb in the
number of refugees from Africa who are Somali after the year 2008. This can be explained by
the 2007 humanitarian crisis, which occurred in Mogadishu displacing nearly 300,000 residents.
According to Menkhaus (2007), in June of 2007, the capital – Mogadishu – was caught in the
grip of heavy fighting between the TFG and a complex insurgency of clan militias, warlords and
Somali Islamists who were seeking more power. This fighting, which was sporadic at times,
lasted for months destroying hospitals, entire neighborhoods and infusing the global diaspora
with a new flow of Somali refugees.

Somali Refugee Resettlement in the United States
Since the early 1990s, the Somali have been coming to America in fairly large numbers
as a result of the conflict between warring factions in Somalia. Prior to 1990, there were
relatively few Somali in the United States – less than 2300. By the year 2000, the U.S. Census
placed the Somali population count at 36,313. Surprisingly, compilations of data from the
Department of Homeland Security and the Office of Refugee Resettlement, reveals a different
picture. The number of Somali refugees, asylees, and non-immigrants admitted to the United
States between 1990 and 1999 alone was 36,693. There are a number of reasons for the
conflicting accounts of the number of Somali in the United States. Some believe that the Somali
are undercounted in the census because they live in inner-city neighborhoods that are often
avoided by census employees out of fear. In addition, the Somali are fearful of government

47

officials because of past experiences in Somalia, so they avoid government representatives
whenever possible (Goza 2007). Whatever their reasons maybe, given the multiple categories
under which Somalis gain entrance into the United States and the inherent shortfalls in census
coverage of this population, any count of the Somali in this country is sure to be conservative.
Despite the discrepancies in count, there is consensus that Columbus, OH and Minneapolis, MN
have the largest concentrations of Somalis in the United States (Kusow 2006). Figure 2.6 and
Table 2.3 present a graphical and numerical representation of Somali admissions from 1990
through 2009. Each peak and trough in the graph, and the corresponding numerical data from
the table, can be traced to events at both the state and international scales with respect to the
number of Somali admissions.

Figure 2.6: Somali admissions 1990 - 2009.
SOMALI ADMISSION TOTALS 1990 - 2009
16,000
14,000
12,000
10,000
8,000
6,000
4,000
2,000
0
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
SOMALI

48

Table 2.3: Somali admissions by category 1990 - 2009.
Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Total:
Percent of
Total:
AVG
(Year):

Refugees

Asylees

I-94

Total

52
192
1,570
2,753
3,555
2,506
6,436
4,974
2,951
4,320
6,026
4,951
237
1,994
13,331
10,405
10,357
6,969
2,523
4,189
90,291

204
117
122
121
150
286
529
708
1,310
2,312
2,404
1,633
667
291
235
162
163
180
169
254
12,017

634
891
1,025
1,004
441
289
383
328
223
271
248
243
5,980

256
309
1,692
2,874
3,705
2,792
6,965
5,682
4,895
7,523
9,455
7,588
1,345
2,574
13,949
10,895
10,743
7,420
2,940
4,686
108,288

83.4%

11.1%

5.5%

100.0%

4514.6

600.9

498.3

Cumulative
Total
256
565
2,257
5,131
8,836
11,628
18,593
24,275
29,170
36,693
46,148
53,736
55,081
57,655
71,604
82,499
93,242
100,662
103,602
108,288

5414.4

Source: Data compiled from Department of Homeland Security and the
Office of Refugee Resettlement special tabulation.

49

Between 1990 and 2009, over 108,000 Somalis were admitted to the United States. Of
those who entered, 83.4 percent were admitted as refugees, 11.1 percent were admitted as
asylees, and 5.5 percent were admitted as I-94 non-immigrants. On average, 5414.4 Somalis
were admitted each year. The Somali who were resettled in the United States via the refugee
resettlement program between 1990 and 2009 were resettled in over 47 states. Five of the 47
states accounted for 46.8 percent of their placement – Minnesota (18 percent), California (8.3
percent), Georgia (7.1 percent), Texas (6.8 percent), and Ohio (6.6 percent). On a smaller
geographic scale, the Somali have been resettled in over five hundred cities between 1990 and
2009 (Figure 2.7), but four cities in particular account for the largest percentage of the placement
– Minneapolis MN (9.9 percent), San Diego CA (5.7 percent), Columbus OH (5.5 percent), and
Seattle WA (3.2 percent).

The major clan-families in Somalia – in addition to being different in subtle linguistic and
cultural ways – differed in proportionality. The Darood constituted 22 percent of the population,
the Isaaq constituted 17 percent, the Dir constituted 7 percent, the Hawiye constituted 25 percent,
and the Rahanweyn constituted 17 percent – the remainder of the population was composed of
the minority (Akiwumi and Estaville 2009). The resettled Somali population within the United
States, however, reveals a distribution that is quite different (Table 2.4). The Somali groups
considered as minorities in Somalia constituted the majority of the Somalis resettled in the
United States. The term ―minority,‖ as used by both Somalis and non-Somalis, is said to refer to
any clan or community that does not belong to one of the four major clan-families that are
considered to be ―noble‖ – the Darood, the Hawiye, the Isaaq, and the Dir (Immigration and
Refugee Board of Canada 1995).

50

Figure 2.7: Somali distribution according to the ORR initial community of resettlement.

51

Out of the major clan-families, as identified by the individual refugees on admission, the
Darood constituted the largest percentage making up close to one-fourth of the Somali that were
resettled in the U.S. between 1990 and 2009. The remaining major clan-families combined
constituted less than 3 percent of the resettled Somali.

Table 2.4: Resettlement figures by clan-family membership 1990 - 2009.
Majority Groups
Darood
Hawiye
Rahanweyn
Issaq
Dir
Total:

% of Resettled
21.28 %
1.22 %
0.71 %
0.63 %
0.26 %
24.1 %

Minority Groups
Benadir
Midgan
Others
Unknown
Total:

% of Resettled
7.13 %
1.31 %
59.5 %
7.97 %
75.9 %
100.0 %

The geography of the Somalis in the United States is rapidly changing. According to
Schaid and Grossman (2007), Somalis are rapidly diffusing out from larger metropolitan areas in
secondary migrations, and settling in smaller frontier communities. The establishment of the
Somali population in these new communities is thus reinforced through affidavits of relationship
which spurs ORR considerations of the community as a new site for refugee resettlement. This
new ORR designation for the community, however, may take years to establish.

Corruption in the Refugee Priority 3 Program
Refugees are resettled in the United States in accordance with three priorities. The first
two priorities address those persons or groups at imminent risk and those persons or groups of
special interest or humanitarian concern to the United States. The third priority (P-3) is for

52

persons identified for family reunification. In the summer of 2008, the P-3 program was
suspended due to widespread fraud ―when U.S. officials found that most refugees from Africa
using the P-3 program were not related at all. The fraud rate among Somali refugees was
reported to be as high as 90 percent‖ (Barnett 2010). As part of a pilot program, The State
Department‘s Bureau of Population, Refugees, and Migration (PRM) and The Department of
Homeland Security (DHS) began DNA testing of three hundred applicants in response to
allegations of fraud within the program. The pilot test revealed a high rate of fraud which
facilitated an expansion of the testing to incorporate more countries in Africa – where more than
95 percent of the P-3 applicants reside. Most of the three thousand P-3 applicants in the
expanded tests came from Somalia, Ethiopia, and Liberia (Bureau of Population, Refugees, and
Migration 2009). Based on the results of the expanded testing, all applications for the program
were stopped in order to allow time for the development and implementation of new procedures
to verify family relationship claims. To this researcher‘s knowledge, the family reunification
program has yet to be restarted.

American Concerns about Somalis
There could be many reasons for the issues we call ―fraud‖ within the family
reunification program. One reason could be purely based on the United States Government‘s
concept of what constitutes a family and its inherent link with DNA. While the United States
Government‘s notion of immediate family is still heavily influenced by biological association,
persons from a war zone might conceive family simply as those who have taken care of them for
the largest portion of their life without regards to biology. There have been many instances
where children, who just lost their parents in the carnage, were picked up by other adults who

53

then took care of them like their own. No matter what the cause, however, the fraud issue is just
one more event that feeds into the negative image some of the more racist elements in American
society try to use in their portrayal of the Somali immigrant. As with changes in immigration
policies after 196520, refugee policy changes under the 1980 Refugee Act expanded the regions
from which refugees were accepted. The opponents of refugee resettlement adhere to the same
argument cast against other non-European immigrants. They argue that the new admits are
drastically different from the primarily ―European stock of previous periods‖ and they have little
chance of assimilating (Frazier and Margai 2003, 25).

Somali Americans Linked to Terrorism
There are other issues pertaining to the Somali that have raised American concerns about
their admission to the United States. The most pressing concern was the Somali link to terrorism.
There was growing concern about the role of Al Itihad al Islaami in sustaining the Al Qaeda
network. In 2001, it was reported that the Al Qaeda and Al Itihad were operating in Somalia
with 3,000 to 5,000 members and 50,000 to 60,000 supporters and reservists (Le Sage 2001). Al
Itihad was in the process of setting up an Islamic state in the 1990s – the movement was linked
(unconfirmed) to the killing of 18 U.S. soldiers during the United Nations Operation in Somalia
(UNOSOM) intervention in Mogadishu in 1993 (p.472). In 2008, dozens of young male Somalis
in the Minneapolis area disappeared causing community members and government officials to
suspect they were returning to the Somalia region to join terrorist groups. This fear was
20

The 1965 Immigration and Nationality Act – also known as the Hart-Celler Act – effectively
eliminated the prejudicial legislation of the 1952 Walter McCarran Act. The abolition of the
quota and preference systems, and the labor certifications led to clear changes in the origin of
immigrants as more and more immigrants from non-European countries – like Asia, South
America, and the Caribbean – gained entry (Keely 1971).
54

confirmed when one of the missing – a naturalized U.S. citizen and graduate of one of the local
high schools – died in a suicide bombing in northern Somalia (Gartenstein-Ross 2008). In more
recent news, in Portland Oregon, a Somali-born teenager plotted a bombing of a Christmas tree
lighting ceremony in the attempt to create a ―spectacular show‖ of terrorism (New York Post
2010). Even though the attempt was unsuccessful, it reinforced negative stereotypes of Somali
immigrants thus polarizing the view of the inhabitants of the communities in which they are
resettled.

Somali Pirates
Another issue that concerns Americans is the prevalence of piracy occurring off the
shores of Somalia.

According to Elliot Anderson (2010), the international community is

polarized by the clash between those persons trying to survive by fulfilling their basic needs and
those who pursue self-serving ends through almost any means necessary. This polarization is in
regards to the act of piracy, and piracy ―is the fastest growing industry in Somalia and is
projected to have replaced agriculture and fishing as the country‘s largest generator of revenue‖
(p.323). In 1994, the U.N. Convention on the Law of the Sea created zones in which coastal
countries have exclusive rights to the natural resources within their zone. For the Somali, this
was a blessing considering their coastal waters were rich with tuna. However, after the collapse
of the Somali government, foreign fishing ships moved in and depleted the reserve of fish in
Somali waters. Some assert that, as a result of these actions, Somali fishermen were forced to
take matters into their own hands by arming themselves and confronting the ―illegal commercial
fleets, demanding they pay taxes, and attempting to drive the vessels from their territory‖
(p.327). In 2010, fifty-three ships were hijacked worldwide with forty-nine of them occurring

55

off the coast of Somalia – the Somali pirates took a record 1,181 hostages receiving millions in
ransom (BBC News 2011). According to Anderson, the stock depletion of Somali water by
foreign fishermen forced the Somalis to adapt to the economic crisis by utilizing other means to
either drive out the competition or capitalize on their participation. While this sounds like an act
of survival, incidents as of late typify acts that polarize the international community. At the
beginning of 2011, the Somali pirates captured and killed four elderly Americans as they sailed
across the world in their yacht (MSNBC 2011). The act of killing the elderly couple is just one
of many instances where innocent persons became prey to those in Somalia who pursue selfserving ends by any means possible. To many, these acts serve as additional reasons to prevent
the admission of the Somali population. When the incidents of piracy and the proclamation of
terrorist are combined with the knowledge that some of the Somalis have been resettled in this
country under fraudulent reasons, entrance under humanitarian concerns can quickly change to
restriction based on security concerns.

Chapter Summary
Somalia is a clan-based society that has been deeply entrenched in fighting between clans
and clan factions since the early 1990s. Some look at the occupation of Somalia by three
colonial powers as the root of the current problems that plague the country today. However,
there are others that point to issues that were endemic to the society prior to colonization that are
just as prevalent today. It is these issues that continuously fuel the on-going conflict. Fighting
between (and within) clans were a part of life prior to the occupation – but not immediately after
the occupation. At one point in time, after independence, the country came close to being truly
united as a nation as pan-Somalism took hold. Somalia experienced a feeling of unity that even

56

surpassed the unity created by the anti-colonial jihad led by Sayyid Muhamed Abdallah Hassan –
referred to as the ―Mad Mullah‖ by the British – at the beginning of the 1900s. The Dervish
movement that began as a pan-clan crusade against colonial forces however, soon degenerated
into intra-clan warfare (Adam 2008).

In an almost instant replay of the theme surrounding the Dervish movement, the unity
garnered by the Barre regime quickly turned to strife as fellow countrymen discovered that while
General Barre stood on an anti-clan platform, he secretly stacked high ranking government
position with clan and relative clan members shifting the balance of power and the distribution of
resources in the favor of his own. Barres actions facilitated a path that ultimately led to all out
civil war between clan and clan factions that sent and continues to send tens of thousands of
refugees across the world in a global diaspora. Over a dozen reconciliation conferences have
been held at international behest since the 1990s with little resolve in uniting the country in what
many believe to be issues of representation. That is, representation for all groups in Somalia. A
solution for this issue was attempted with the adoption of the ―4.5‖ formula, however, the
existence of the ―.5‖ meant that some would be underrepresented. Thus the fighting continues
amidst the splintering of clans and shifting allegiances. In the middle of the chaos, the region
has managed to produce terrorists and pirates. The flourishing of these two elements has not
gone unnoticed by the international community, and their existence in Somalia has caused some
to question the commitment to the resettlement of the Somali people.

57

CHAPTER 3: REFUGEE AND REFUGEE RESETTLEMENT

Refugees
For well over a century, the United States has had a guarded gate – this gate was open to
some, closed to others. The bias of admissions reflected a convergence of factors ranging from
prejudice, public opinion, and political agendas to the economy and administration personalities.
Thus, immigration policy has been ―dominated not by the will to admit, but by the will to
restrict‖ (Zucker and Zucker 1987, 2). Restrictionist sentiment in immigration law had its
beginnings in the mid-1800s with the Page Act of 1875 which was closely followed by The
Chinese Exclusion Act of 1882. The restrictions of the early 20th century aimed at traditional
immigrants also affected the acceptance of refugees because there was no distinction between
refugee and immigrants prior to World War II – even though the number of refugees worldwide
had increased exponentially as a result of the 1917 Soviet Revolution. One of the predominant
factors that kept refugees from even being discussed, as a topic for admission considerations,
rested in the ―likely to become a public charge‖ (LPC) provisions (Zucker and Zucker 1987).
The LPC clause was conceived to restrict the immigration of those who had little in the way of
capital – refugees, given their predicament, were practically guaranteed to become a public
charge at one point or another during the resettlement process. Thus, refugees were not looked at
as people who have lost all the elements that contribute a sense of self-worth, or as people who
have lost the legal and political protection of a government (Zucker and Zucker 1987); rather,
they were looked at as immigrants who would become a drain on the American economy. This
view of refugees still holds today. Immigrants are seen as occupying a ―precarious‖ position
because they must demonstrate they are capable of self-support; refugees, in contrast, are entitled

58

to cash assistance and other benefits and services that are in many ways equal to that of citizens
(Gold 1992).

Pre-1980 Refugee Legislation
Official recognition of differences between refugees and immigrants with respect to U.S.
immigration law did not occur until well after the middle mark of the 20th century. In the 1930s,
a variety of bills were introduced by refugee advocates in an attempt aimed at helping German
refugees; however, restrictionist ideology with respect to immigrants was still the dominant
ideology. The only legislation of promise at that time was the Wagner-Rogers bill which
allowed for the entrance of German refugee children on a non-quota basis. About a decade later,
the first significant piece of legislation pertaining to refugees was introduced – the Displaced
Persons Act of 194821. Despite the legislation being laced with restrictionist sentiment, it still
reflected humanitarian concerns with refugees that were never present in prior legislation – the
act was amended in 1950 and 1951 culminating in the admission of over four hundred thousand
refugees (Zucker and Zucker 1987). The giant step forward, however, was preceded by a giant
step back as the McCarran-Walter Immigration and Nationality Act of 1952 was passed.
McCarran‘s argument in defending the act rested on racist notions – born out of the scientific
racism which was heavily present in the early 1900s – that was used to draft the original quota
acts of the 1920s.

21

Displaced persons (DP) was a term used to denote the survivors of the Holocaust and others
who were uprooted by war, or who had been expelled or fled from their homeland with no hope
for return.
59

Challenges to the restrictionist mentality multiplied in the early 1950s in response to an
escalation in Cold War activity; first there was the Refugee Relief Act of 1953, which relaxed
the restrictions set by the McCarran-Walter Act. Next there was the Refugee Fair Share law
whose subsequent extensions eventually led to the incorporation of refugee admissions into
immigration law. Under the law, a refugee was continuously defined as ―an individual fleeing
racial, religious, or political persecution from any Communist, Communist-dominated, or Middle
Eastern country‖ (Zucker and Zucker 1987). Thus, this definition resulted in an association
between refugees and communism. As a result, refugee admission was now seen as thwarting the
spread of communism. The communistic element in the U.S. definition of refugee resulted in
waves of Soviet Jews, Cuban, and Indochinese admissions – each with their own special
program. The creation of the ―Refugee Act of 1980 was intended to replace the patchwork of
varying programs that had developed in response to the individual refugee crises‖ (Zucker and
Zucker 1987, 50).

The Refugee Act of 1980
Refugees are subject to the Refugee Act of 1980, which contains policies relating to their
resettlement based on government objectives of integration, expedient assimilation, and selfsufficiency. It was not since 1952 that the laws on admission and resettlement of refugees had
been fundamentally reformed; with the 1980 act, Congress gave new statutory authority to the
U.S. commitment to human rights and its humanitarian concern for the plight of refugees around
the world (Kennedy 1981). The Refugee Act of 198022, which incorporated into American law

22

Provisions for the Refugee Act are contained within Title IV, chapter 2 of the Immigration and
Nationality Act (INA).
60

the international definition of a refugee, was passed by Congress to provide a uniform procedure
for the admission of refugees and to authorize federal assistance for their resettlement and
procurement of self-sufficiency (Bruno 2008). Uniformity was necessary because from the end
of World War II up until 1980, refugee admissions were always under special legislation, or
some strained interpretation of the parole authority (Zucker and Zucker 1987). The Refugee Act
of 1980 accomplished six basic goals (Kennedy 1981, 143):
1. It repealed the previous law‘s discriminatory treatment of refugees by providing a
new definition of refugee. The new definition no longer applied only to refugees
―from communism‖ or certain areas of the Middle East; it applied to all who met the
test of the United Nations Convention and Protocol on the Status of Refugees.
2. It raised the annual limitation on regular refugee admissions from 17,400 to 50,000
each fiscal year.
3. It provided for an orderly but flexible procedure to deal with emergencies if refugees
of ―special humanitarian concern‖ cannot be resettled within the regular ceiling.
4. It replaced the use of ―parole authority‖ with new statutory language asserting
Congressional control over the entire process of admitting refugees.
5. It established an explicit asylum provision in the immigration law for the first time.
6. It provided a full range of federal programs to assist in the resettlement process and
created the Office of the United States Coordinator for Refugee Affairs and the Office
of Refugee Resettlement to monitor, coordinate and implement refugee resettlement
programs.
A refugee, thus with this new act, was defined as a ―person who is outside his or her
country and who is unable or unwilling to return because of persecution or a well-founded fear of
persecution on account of race, religion, nationality, membership in a particular social group, or
political opinion‖ (Bruno 2008, 1).

61

Refugee Resettlement
In 2006, there were over 12 million refugees in the world (Singer and Wilson 2006). For
these refugees, the U.N. recognizes three durable solutions23: voluntary repatriation to their
homeland; integration into host society; or resettlement to a third country (Stein 1983). Out of
the 22 resettlement countries (Figure 3.1), the United States accepts more refugees than the other
resettlement countries combined (Patrick 2004).

The response of host countries to this

increasing population varies widely between states and governments – responses also vary over
time and by refugee group (Jacobsen 1996).

United States Refugee Policy
An important goal of U.S. foreign policy, ever since the government established a firm
policy for refugees, is their continued assistance and resettlement – since 1980, more than 2.5
million have been resettled in the United States (Figure 3.2). As evident in Figure 3.2, the
largest count of annual refugee admissions occurred immediately after the enactment of the 1980
Refugee Act. The annual number of refugees admitted eventually decreased and have never
again reached the initial levels seen in the early 1980s. The lowest count of refugee admissions
occurred immediately after the terrorist attack on 9/11; since that time, the number of refugees
admitted has been slowly recovering. It was not until the year 2009 that the level of refugee
admissions to the United States reached the pre-9/11 attack levels.

23

Durable solution means helping the refugees to become self-sufficient, enabling them to
integrate and participate fully in the social and economic life of their new country (Stein 1983,
190)
62

Figure 3.1: The 22 countries actively participating in the resettlement of refugee.

63

Figure 3.2: Graphic depiction of refugee admission 1980 - 2009.

REFUGEE ADMISSIONS - 1980 to 2009
250,000
200,000
150,000

100,000
50,000
0
2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

REFUGEES

Refugees are processed and admitted to the United States abroad based on a system of
three priorities for admission (Bruno 2009):
Priority One: Persons for whom no durable solution exists who are referred to the U.S.
refugee program by the United Nations High Commissioner for Refugees (UNHCR), a
U.S. embassy, or designated non-governmental organization (NGO) because of imminent
risk.
Priority Two: Groups of special interest and humanitarian concern to the United States.
Priority Three: Persons identified for family reunification; they can by cases of spouses,
unmarried children under 21, and parents of persons who were admitted to the U.S. as
refugees or granted asylees.
The U.S. State Department handles the processing of refugees and the Department of Homeland
Security24 (DHS) makes the final determination on the eligibility of the refugees for admission;

24

Immigration and Naturalization Service (INS) previously handled refugee adjudications;
however, the DHS abolished the INS in 2003 and assumed its functions.
64

after one year in refugee status in the U.S., refugees are required by law to apply to adjust their
status to legal permanent resident (Bruno 2008). The number of refugees admitted to the U.S.
each year is determined by the President in consultation with Congress; according to Bruno
(2008), the President submits a consultation document containing a proposed refugee ceiling and
regional allocations for the ensuing fiscal year (Figure 3.3).

Figure 3.3: Refugee admissions and ceilings.

REFUGEE CEILING AND ADMISSIONS 1990 - 2009
160,000
140,000
120,000
100,000
80,000
60,000
40,000
20,000
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
CEILINGS

ADMISSIONS

The regions allocated for refugee resettlement are: Africa, East Asia, Europe and Central Asia,
Latin America/Caribbean, and Near East/South Asia. Along with the regional allocations, there
is also an Unallocated Reserve category in case an event arises that would require the need for
additional admissions.

65

The refugee ceilings represent the maximum limit for refugee admissions alone and have
no bearing on the number of persons granted asylum; for asylum seekers, there are no numerical
limits. For refugee admission between 1990 and 2009, the average ceiling was 95,588 – the
admissions during this period were averaged at 80.9 percent of the ceilings (Table 3.1). The
remaining percentage of the ceiling that is not filled by admissions is held for those who could
become potential admissions; this unallocated reserve is used if, and where, a need develops for
refugee admissions in excess of the allocated numbers (Bruno 2008).

Refugee policy set forth by the 1980 Refugee Act dictates the dispersal of refugees – the
head of the Office of Refugee Resettlement (Director) shall:
… insure that a refugee is not initially placed or resettled in an area highly impacted (as
determined under regulations prescribed by the Director after consultation with such
agencies and governments) by the presence of refugees or comparable populations unless
the refugee has a spouse, parent, sibling, son, or daughter residing in that area. (Office of
Refugee Resettlement 2010)

Even though refugees have been settled throughout the United States, ten states in particular
received the vast majority of them (Table 3.2). Singer and Wilson‘s (2006) examination of the
refugees resettled between 1983 and 2004 revealed that California received the majority – 25
percent. California was closely followed by New York which received 15 percent. The data
reflects the states of initial resettlement, thus secondary migration would necessarily change
those figures. Much of the secondary migration occurs within the first few years after arrival and
then the refugee group becomes relatively stabilized in their spatial distribution after an initial
period of adjustment (Office of Refugee Resettlement 2006).

66

Table 3.1: Refugee Admissions and Ceilings 1990 - 2009.
Year

Ceiling

Admits

2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990

80,000
80,000
70,000
70,000
70,000
70,000
70,000
70,000
80,000
90,000
91,000
83,000
78,000
90,000
112,000
121,000
132,000
142,000
131,000
125,000

74,602
60,108
48,281
41,279
53,813
52,840
29,320
26,996
68,388
72,519
85,014
76,750
76,456
75,755
99,553
112,065
119,050
131,749
113,980
122,935

% of
Ceiling
93.3%
75.1%
69.0%
59.0%
76.9%
75.5%
41.9%
38.6%
85.5%
80.6%
93.4%
92.5%
98.0%
84.2%
88.9%
92.6%
90.2%
92.8%
87.0%
98.3%

AVG:

95,588

79,910

Cumulative
Admissions
1,541,453
1,466,851
1,406,743
1,358,462
1,317,183
1,263,370
1,210,530
1,181,210
1,154,214
1,085,826
1,013,307
928,293
851,543
775,087
699,332
599,779
487,714
368,664
236,915
122,935

80.9%

Source: ORR Report to the Congress FY 2006; CRS Report
RL31269, 2009.

67

Table 3.2: Top ten refugee resettlement states and totals
1983 - 2004.
Rank State
1
California
2
New York
3
Texas
4
Washington
5
Florida
6
Illinois
7
Massachusetts
8
Pennsylvania
9
Minnesota
10
Georgia
Total Refugees:

Number of
Refugees
405,806
235,325
85,750
81,857
73,211
70,248
54,000
52,095
48,820
43,068
1,655,406

% of All U.S.
Refugees
24.50%
14.20%
5.20%
4.90%
4.40%
4.20%
3.30%
3.10%
2.90%
2.60%
69.30%

(Singer and Wilson 2006, 8)

A precise timetable for the duration of the initial adjustment period does not exist because the
amount of time would vary for each refugee group. Again, it has been suggested that better
employment opportunities, established ethnic communities, better welfare benefits, family
reunification, or more congenial climates serve as pull factors luring refugees away from their
initial sites of resettlement (Office of Refugee Resettlement 2008).

Changes in Refugee Policy
In accordance with the 1980 Refugee Act, states were eligible for 100 percent
reimbursement for the services they provided to the refugees for the first three years of the

68

refugee‘s residence in the United States (Stein 1983). The three years of assistance was assumed
to be the amount of time it took for refugees to become fully immersed in American society –
economically and culturally. According to Zucker (1983), services provided to refugees came
under fire when Ronald Reagan took office soon after the initiation of the 1980 Refugee Act.
The Reagan administration executed an ambitious policy of social program reduction – federal
funding for social services like job training, education, and welfare were cut or eliminated. The
commitment to refugee resettlement was reduced by one-half forcing the limitation of services
from 36 months to 18 months and the limitation of the reimbursement to states from 100 percent
to 50 percent. The restrictive approach of the Reagan Administration towards the resettlement
of refugees and the controversy surrounding the approach called attention to the problems
inherent in the resettlement program (Zucker 1983). Ensuing changes to refugee policy has since
cut the time limit set on services offered to refugees from 18 months to 4 – 8 months. Refugees,
however, are eligible for other programs (Table 3.3) which operate on different time limits
(Bruno 2009).

Table 3.3: Refugee eligibility timeline for major federal assistance programs.
Program:
Supplemental Security Income
Medicaid
Temporary Assistance for Needy
Families
Food Stamps

Eligibility Time Limits:
7 years after entry
7 years after entry, then state option
5 years after entry, then state option
No time limits

(Bruno, Congressional Research Services: Refugee Admissions and Resettlement
Policy 2009)

69

Refugee Model and a Refugee Type
Based on the amount of refugees accepted into a host country and the country‘s agenda
for incorporating those refugees, models can be made of the refugee resettlement process – the
United States adheres to the Large Volume / Primacy on Economic Adaptation Model of refugee
resettlement (Lanphier 1983). In this model, refugee resettlement is seen as an activity of
sponsors from the private sector and VOLAGS, which – according to Lanphier (1983) – aspire to
provide goods and services to refugees on a short-term basis in order that they may assume full
engagement as productive members in the labor force in their new host country.

In a

contemporary sense, this task has become more daunting because of a shift in the receiving
regions. Historically, the United States has received refugee groups from developed regions of
the world where differences in skill and culture were nominal. Now, a growing proportion of
refugees are coming from underdeveloped parts of the world or from areas of the world where
statelessness (as with the Somali) prevails – transferable skills and cultural similarities are
basically non-existent. These types of refugees require a greater amount of assistance because it
is typically harder for them to adapt to American life due to the contrast in culture. The changes
instituted during the Reagan administration compounded this situation not only by limiting the
ability of resettlement agencies to do their jobs adequately, but it also placed a greater amount of
pressure on the refugees themselves to quickly conform.

Table 3.4 list figures from the originating regions for refugee groups resettled in the
United States between 1946 and 2000 – the vast majority of refugees are still coming from places
such as Europe and Asia, but a growing number are coming from Africa.

70

Table 3.4: Refugees LPR by region of origin 1946 - 2000.
1946 to 1951 to 1961 to 1971 to
1950
1960
1970
1980
Europe
211,983 456,146 55,235 71,858
Asia
1,106
33,422 19,895 210,683
Africa
20
1,768
5,486
2,991
*North America
163
831 132,068 252,633
* North America includes the Caribbean and Central America
Region

1981 to
1990
155,512
712,092
22,149
121,840

1991 to
2000
426,565
351,347
51,649
185,333

Source: U.S. Department of Homeland Security

Despite the African region showing the least amount of refugees resettled over the five decade
period from 1950 to 2000, there was a 640 percent increase in numbers applying for legal
permanent residence25 during the 1980s, and a 133 percent increase over that figure applying for
legal permanent residence during the 1990s.

The growth in African refugees, however,

corresponds with a growth in concerns over the increased cost and difficulties of resettlement –
particularly with third-world refugees. This concern – in combination with the endurance of the
shortened amount of time for assistance – has led some observers to wonder if the commitment
to refugee resettlement as a durable solution has been diminished (Stein 1983).

Voluntary Agencies (VOLAGS)
In nearly all of the countries that open their doors for refugee resettlement, the
resettlement process is accomplished through partnerships between the country‘s government

25

Refugees are eligible to apply for legal permanent residence one year after initial resettlement
in the United States.
71

and VOLAGS. VOLAGS are national organizations and within the U.S., ten VOLAGS (five of
which are faith-based) carry out most of the refugee resettlement (Table 3.5).

Table 3.5: Refugee resettlement voluntary agencies.
1. Church World Service (CWS)
2. Ethiopian Community Development Council (ECDC)
3. Episcopal Migration Ministries (EMM)
4. Hebrew Immigrant Aid Society (HIAS)
5. International Rescue Committee (IRC)
6. Kurdish Human Rights Watch, Inc. (KHRW)
7. US Committee for Refugees and Immigrants (USCRI)
8. Lutheran Immigration and Refugee Services (LIRS)
9. United States Conference of Catholic Bishops (USCCB)
10. World Relief Corporation (WR)

The VOLAGS have a network of nearly four hundred affiliates across the U.S. dedicated to the
reception and integration of refugees. VOLAGS constitute a key element in the resettlement
process; they are contracted with – and partially funded by – the Bureau of Population, Refugees
and Migration (PRM), and are responsible for providing reception and placement services for
refugees upon their arrival in the U.S. (California Department of Social Services 2008). While
still overseas, applicants considered for resettlement are interviewed and screened by the U.S.
State Department and the Department of Homeland Security. Once approved for admission to
the United States, the refugees are allocated among the 10 national VOLAGS. Each year, these
VOLAGS propose the resettlement communities and the number of refugees the communities
will receive – the proposals are subject to approval by the U.S. State Department. The local
affiliates of the national VOLAGS that operate throughout the country receive the refugee
referrals from their parent agency and coordinate their resettlement within the local communities.

72

The VOLAG-Government Partnership
The VOLAG partnership with the federal government initially started with the 1946
Corporate Affidavit Program. According to Zucker (1983), under this program, the VOLAGS
guaranteed to provide sole financial support to refugees from self-generated resources to prevent
them from becoming public charges. Since most of the displaced persons had no means of
financial support, the affidavit effectively enabled the entrance of thousands of refugees into this
country who otherwise would have never had the chance. The financial relationship between the
federal government and the VOLAGS changed after World War II when the government first
began extending loans to the agencies which were initially for refugee transportation costs
(Wright 1981).

Nearly two decades after the initial establishment of the un-funded partnership, the nature
of the relationship changed as the government began issuing more loans in response to the
increasing post-World War II refugee flow.

The amounts of the loans, over time, were

incremental and differential for the various refugee groups:
Small per capita resettlement grants were made for the first time with the Cuban refugee
movement through the 1960s. The grants reached $300 per capita; a quantum jump, for
the small Ugandan Asian program in the early 1970s, and $500 per capita in the initial
Indochinese program. Grants for the resettlement of Indochinese refugees dropped to
$350 and rose again in stages to the current $500 per capita. Until 1977, no grants were
extended for the resettlement of the regular flow of non-Southeast Asian, non-Cuban
refugees. However, in 1976, to provide more equal treatment for refugee groups and to
help the VOLAGS with rising numbers and costs, the State Department began granting
the VOLAGS $250 per capita for at least some part of the annual refugee admissions and
later raised that to $350. (Wright 1981, 161)

73

Thus refugee sponsorship was initially handled by private parties but the 1980 Refugee Act
solidified the promise of government funding and from that period till the present, public funds
have dominated all aspects of refugee resettlement (Barnett 1999).

Problems in Refugee Resettlement
For the 2011 fiscal year, the recommended per capita funding for VOLAG reception and
placement of refugees was $1800. According to Zucker (1983), there were severe deficiencies in
the resettlement program because VOLAGs operate under loose financial requirements and a
loose definition of core services. Special concern was directed at how funding was being used
and how the figures for funding were derived for the resettlement of refugees. According to
resettlement watch dog groups, cost-analysis reports from VOLAGs are troublesome if those
reports are inaccurate. The most recent proposal for increased government funding has raised
questions as to the exact percentage VOLAGs contribute to the resettlement process. According
to Coen (2010), agencies claim that funding for the resettlement costs are divided up as such
among the contributing entities: 39 percent government funds, 30 percent private party funds,
and 31 percent VOLAG funds. Issues of validity of the proportion of funding arise when the
VOLAGs cannot account for the 31 percent of the costs for refugee resettlement. In addition,
more questions surface concerning the contribution of the private entities. Although those in the
private sector do indeed contribute a small percentage, audits would appear to put that
contribution around 8 percent rather than the 30 percent as claimed by VOLAG reports (Coen
2010). This translates into the government assuming an increased financial role in the reception
and placement of refugees.

74

When the increased federal government funding of refugees per capita and the large
number of refugees the United States admits are combined – the role of the VOLAG becomes
more business oriented. In such an environment it is easy to equate more refugees with more
money and in the United States – most refugees are assigned to one of ten VOLAGs adding to
that agency‘s headcount and thus their federal cash allotment (Barnett 1999). This scenario
becomes a problem when one considers the primary origin of contemporary refugees. As stated
earlier, more and more refugees are coming from developing countries – this poses a challenge to
the VOLAG workers because practices that worked with European refugees are inappropriate for
use with the newest arrivals (Stein 1983). Prior refugees came from European countries in which
the culture was fairly similar to the culture of this country thus the constraints on acculturation
were limited. Because the culture of most of the contemporary refugees vary drastically from
that of this country, more time and effort is needed in providing services making their
resettlement very difficult and very expensive (Stein 1983). With a decrease in the time allotted
for refugee support and an increase in the number of refugees being admitted who require a
longer period of support, Stein‘s supposition of decreased commitment to refugee resettlement
becomes more of a reality.

The Welfare System
When considering the changes that have occurred to refugee policy, one primary question
that arises is – how are the VOLAGS able to do more with less when preparing refugees to
become self-sufficient in the quickest amount of time possible? According to Wright (1981),
when the influx of refugees tax the resettlement system, VOLAGS resort to temporarily placing
them on public assistance until they can cope with the newcomers.

75

Under extreme

circumstances, the ―welfare programs are necessary and beneficial if properly used,‖ (p.163) –
but the existence of such a safety net can lead to abuse by the VOLAGS. Don Barnett (1999)
believes this is exactly what happened as he asserts that government money corrupted the
refugee resettlement process. In a documentation of one of the largest ORR conferences in
history, Barnett described the event as a gathering in which voluntary agency workers learned
―how to qualify for refugee-specific grant money‖ in addition to learning how to ―tap into much
broader federal initiatives‖ (Barnett 1999). The conclusion drawn from observing the conference
sessions led Barnett to believe that their main purpose was for the dissemination of information
on how to get refugees into the welfare system as quickly as possible.

The generous welfare system in the United States can also lead to welfare abuse on
behalf of the refugees as they bypass the self-sufficiency goal, and advance straight to welfare
dependency. According to Barnett (1999), a 1996 federal study of refugees that had arrived five
years prior found that 46 percent were receiving cash assistance, 48 percent were receiving food
stamps, and 24 percent were in – or waiting for – public housing. Barnett found this astonishing
given that anywhere from one-third to four-fifths were admitted as family reunification cases –
the family, as sponsors, were supposed to be self-sufficient before accepting responsibility for
the new admits.

Once the government became involved with funding the resettlement of

refugees the process became compromised. The introduction of federal money removed the
impetus for refugees to integrate into American society – with the immediate availability of cash,
food, housing and healthcare, there is no need (and probably no desire) or incentive to do so
(Barnett 1999).

76

One of the direct distinctions between a refugee‘s and an immigrant‘s adaptation to life in
the United States involves the relationship with the welfare system. There is still scholarly
debate as to whether resettlement policies encourage self-sufficiency or welfare dependency –
the availability of public assistance and social services provide refugees with an alternative to
entering mainstream labor markets (Hein 1993). Since the start of refugee admissions, millions
have come to the United States and most have integrated themselves into the American economy
and society at different speeds. At the same time, billions in government funding was invested in
programs geared towards the speedy integration of refugees into the economy, yet, there is still
no evidence to suggest that the money – or the manner in which it was spent – made any
difference to the rate of refugee employment (Fass 1985).

Refugees and the Welfare Magnet States
The preceding passages establishing the association between refugees and the welfare
system was incorporated because the availability of welfare (or lack of it) affects the speed in
which refugees adapt to and integrate in American society (Barnett 1999). Secondly, variables
associated with welfare serve as powerful pull factors in interstate migration – a concept posited
by the welfare magnet hypothesis. The magnet hypothesis has many aspects, but the major
proposition of concern in this research is the notion that welfare benefits affect the residential
location choices of immigrants – the ―welfare state creates a magnet that influences the migration
decisions of persons in the source countries potentially changing the composition and geographic
location of the immigrant population in the United States in ways that may not be desirable‖
(Borjas 2003, 290). While the hypothesis was originally applied to international migrants, it
later became relevant to interstate migrants of this country. The welfare magnet refers to states

77

or locales whose policies attract poor migrants and/or facilitates the retention of a high
percentage of its own poor (Allard and Danziger 2000).

The Personal Responsibility and Work Opportunity Reconciliation Act
The question of welfare usage among noncitizens – especially in light of the fact that they
have a higher rate of usage than citizens as established by Borjas (2003) – sparked the immigrant
related provisions of the 1996 Personal Responsibility and Work Opportunity Reconciliation Act
(PRWORA). The PRWORA created a new set of guidelines for establishing the eligibility of
immigrants to receive public assistance – it made it harder for them, as noncitizens, to be
approved for welfare programs. While the legislation did make an impact on the behalf of
welfare reform, the impact was limited because of behavioral changes on behalf of the states and
the immigrants themselves.

The states with large immigrant populations reacted to the

PRWORA by offering state level benefits to immigrants negatively affected by the welfare
reform, and immigrants themselves also reacted independently by becoming naturalized in order
to regain access to public assistance (Borjas 2003). Not all states reacted by extending state
benefits, fifteen26 actually instituted proactive measures to curb welfare induced migration.
According to Allard and Danziger (2000), proactive states passed welfare residency requirements
that set, for those migrants who have been in the state for less than twelve months, the level of
benefits equal to that of their last state of residency. These restrictions, however, were found to
be unconstitutional infringements on the right to travel according to the 1969 case Shapiro v.
Thompson, a precedent which was upheld on May 17, 1999, by the Supreme Court in Saenz v.
26

The fifteen states included: California, Georgia, Florida, Illinois, Maryland, Minnesota, North
Dakota, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont,
Washington, and Wisconsin.
78

Roe (Allard and Danziger 2000, 351). While the PRWORA might not have seriously affected
refugee usage of welfare, the residency requirements established by the fifteen states mentioned
above might have had an impact on the secondary migration decisions of refugees had it not been
deemed unconstitutional.

The Saenz v. Roe27 decision effectively held that states cannot

discriminate against citizens based on their length of in-state residency (Davis 1999).

Chapter Summary
The purpose of this chapter was to lay the foundation for the type of institutional
environment the Somalis face upon arrival in America as refugees. Again, this is not to say all
Somalis in America are refugees, however, the majority of their population admitted to the U.S.
entered under refugee status. As it has been shown, the refugee program was initially privately
funded, however, the federal government soon became involved with the partial funding of
refugee resettlement – this joint partnership was solidified with the 1980 Refugee Act. The
initial goal of the VOLAGs was to bring the refugees to self-sufficiency – economically and
socially – as soon as possible; however, this goal has been altered to some degree. Since the
inception of the act, the time allotted for refugee support has been reduced, the number of
refugees requiring more time and support in acculturating to this society has increased, and the
financial backing of resettlement agencies have been reworked in such a manner as to create a
system where more money is awarded for resettling more refugees thus removing the
humanitarian drive and replacing it with one that is financial. Thus the contemporary refugees –

27

Residency laws were challenged as a violation of the ―principle of free interstate migration‖
(right to travel) – a principle that holds that ―United States citizens be free to choose their state of
residence and that state citizens be treated equally regardless of length of residency‖ (Davis
1999, 97).
79

especially those who are coming from cultures vastly different from this host country – more
often than not, are rushed through a system that no longer promotes self-sufficiency, but rather
welfare dependency. The changes in the resettlement process since its inception and the
existence of the welfare state all serve to impact the Somali because both may affect the decision
of the Somali to embark on secondary migrations. The fast track orientation and support service
term contemporary refugees, like the Somali, are subjected to pushes them through the
resettlement process and into the mainstream public with inadequate preparation. This lack of
preparation serves to strengthen the impetus for secondary migrations to established ethnic
communities for not only social and psychological support, but also for safety and security. The
existence of the welfare state – if adequate – can serve to keep them in the enclaves and bolster
the ethnic clustering.

Horst (2006) characterizes the Somali heritage as being governed by a mentality
composed of three elements: a mentality looking for greener pastures; a strong social network
supported by obligations to ensure survival; and the sense of strategic dispersal of family
members and activities as a matter of risk reduction. These elements of the Somali culture
suggest a highly mobile population who would search out the best environment for settlement.
Thus, whether they are in the arid lands of Somalia, or in the temperate regions of the U.S.,
historical evidence from Somali ethnic origins infers that they will seek out locales best suited
for survival.

The existence of established ethnic enclaves, or better welfare benefits and

resources provided by various states would definitely factor into the redistribution of Somalis
across the American landscape in their search for greener pastures – especially if the states of

80

initial resettlement contained barriers or constraints that interfered with their idealized residential
setting.

81

CHAPTER 4: REVIEW OF THE LITERATURE

According to Hein (2006), evidence that the history, culture, and politics of an immigrant
group shapes the way they adapt to America – especially with respect to race and ethnic relations
– continues to mount. Despite this mounting evidence, there has been little change to the
prevailing explanations for immigrant adaption which includes assimilation, ethnic competition,
and mode of incorporation theories. Hein believes the reason why the effects of an immigrants
homeland experience has not stimulated alterations to these theories is because the change would
emphasize that there are cultural differences between the groups. While this acknowledgment of
cultural differences may seem benign, Hein asserts that it is just the opposite because of the two
trends that created a ―theoretical dilemma‖ (p.6). One involved the War on Poverty from the
sixties and the neoconservative policy paradigm of the eighties which ―made cultural
explanations for racial and ethnic inequality a social-science taboo‖ (p.6). The second involved
the postmodernist reinterpretation of culture during the eighties and nineties which ―banished
one of its essential meanings that is particularly appropriate for immigrants: values and norms
internalized through socialization‖ (p.6). Hein asserts that, because of those two trends, theories
pertaining to immigrants continue to avoid placing emphasis on the immigrant‘s culture or other
aspects of their pre-migration experience. As a response, Hein developed his ethnic-origins
theory based on cultural aspects such as kinship norms, politics, and religious values:
Owing to global diversity in history, politics, and culture, each group of emigrants takes
with it a unique world view of what it means to be a member of society. These
homegrown conceptions of societal membership influence how immigrants interpret and
respond to new identities and inequalities in a host society. The concept of ethnic origins
helps us appreciate that immigrants add more than demographic breadth to diversity in
the United States. Immigrants also change the meaning of diversity because they have
distinctive ethnic origins. (Hein 2006, 7)

82

Breakthroughs in the acknowledgement of the immigrant‘s ethnic origins have been
evident as of late. More and more studies are now looking at the immigrant‘s experience and
conceptualization of race and ethnicity prior to migration to examine how they adapt to these
concepts post-migration. According to Thai (2008), variations in the experiences immigrants
had in their country of origin affects their access to institutional power and mobilization. These
are factors that help to explain why immigrant groups develop competing understandings of
personal and institutional discrimination (p.613). The focus on race and ethnic adaptation is
evident in Somali studies as the bulk of the literature is geared towards examining how this
group deals with the black-white racial dichotomy that is still prevalent throughout American
society.

Studies of Somali Adaption
Darboe (2003) conducted an analysis of the Somali in Minnesota to understand the
immigration and assimilation process. After drawing on the historic experiences of African
Americans, Darboe found that the Somali face several challenges to assimilation in Minnesota to
include cultural differences, racism, economic inequalities, and language barriers (p.458). In
Minnesota, these challenges were all bound by the role of race – a concept that the Somali (and
many other African immigrants) have trouble grasping because groups are identified and labeled
based on skin color rather than national origin. Darboe examined three major ideologies of
assimilation (Anglo-conformity theory, the melting pot theory, and theories of cultural pluralism)
placing most of the focus on the ideology of cultural pluralism and its implications that ―various
racial or ethnic groups be allowed to freely express their own culture without suffering prejudice
or discrimination‖ (p.467). Darboe asserted that the on-going debate about new immigrants –

83

like the Somali – and the ethnic diversity they create highlights the conflict between the Angloconformity and cultural pluralism ideologies. The end belief is that the appeal of cultural
pluralism, and the racial and ethnic diversity it spawned, has created fear on behalf of the
dominant white population who believe that immigrant pluralism might transition into
separatism; this in turn stimulates even more racism and nativism (p.470). Darboe concludes by
stating that the strategies the Somali have employed as a counter to racism includes building
coalitions with other ethnic groups not only as a show of solidarity, but also as a way to stave off
the separatist label.

Kusow (2006) also analyzed the racial issue in his study of the Somalis in North
America. He used qualitative methods to examine the Somalis understanding of race and
blackness, as defined in North America, to identify how color-based racial categories are
―contested and redefined with a different social classification system that does not acknowledge
blackness as a meaningful category for social stratification‖ (p.534). In Somalia, skin color was
not the basis for social stratification like in the United States – social stratification, as well as
social identities, in Somalia was primarily determined through clannism (p.541).

So the

underlying focus was on how immigrants from societies that do not employ color-based racial
categories negotiate identities in a society that does. Kusow found that Somalis maintain social
boundaries that not only provide barriers ―with which to shield themselves from discrimination,
but also gives them a necessary cultural framework from which to reverse-stigmatize members
of their host society‖ (p.546). These social barriers would be consistent with a research report on
survey answers from the Somalis living in Minneapolis. According to the report, a minimum of
74 percent indicated that they identify more with the Somali culture than the American culture.

84

In addition, only 15 percent of the Somalis indicated that they wanted to become more a part of
the American culture (Wilder Research Center 2000). This, in effect, means the overwhelming
majority of Somali do not want to become a part of the American culture which feeds into what
Darboe (2003) expressed about the fear that cultural pluralism may transform into separatism.
The maintenance of distance from American culture could be a product of the myth of return
which highlights the immigrant‘s belief and desire for an imminent return to their homeland.
Even though the return rarely occurs, the myth still contributes to the ―group‘s need to defy‖
becoming incorporated into the host society via acts of self exclusion (Cohen and Gold 1997).
The distancing can also be a protective measure – some groups may restrict interaction to
prevent the ―dilution‖ of their customs through exposure to other cultures or other groups (Grady
2003).

Again, the preceding studies fit in with the overall trend in Somali studies by focusing on
issues regarding how cultural, economic, social, and racial context inform Somali identity
formations. These are important issues; however, this research is concerned with the geographic
aspects of the Somali in the United States.

Immigrant Geographic Distributions
Studies of Locational Determinants
There have been a number of studies that have assessed the locational determinants of the
foreign-born population. It is this admission that fueled Newbold‘s (2002) research as he noted
large differences in the literature whereby studies of immigration and settlement overlooked the
status of the immigrants thus referring to the foreign-born in aggregate. Newbold believed the

85

oversight was due to the differences in the philosophic traditions of the researchers investigating
the migration phenomenon:
Although the existing literature has explored the foreign-born population with respect to
national origin and arrival cohort, much less empirical work addresses differences by
admission category. In part, this is due to differential traditions within the research
literature. Qualitative techniques or surveys are more disposed to the analysis of a
fragmented refugee population and are typically the only vehicle through which to
examine this population. Quantitative techniques have tended to draw upon human
capital theory, focusing upon the immigrant population. More importantly, the lack of
available and comparative work is typically due to a deficiency of information relating to
admission category in publicly released data files. (Newbold 2002, 371)
Making the distinction between the categories in which the foreign-born are admitted is
important for geographic research because the motivation for immigrating and the differences in
human capital may eventually exert an influence on the settlement patterns and adjustment
behaviors of the foreign-born within their host society (Newbold 2002). Again, a big part of the
problem in making comparisons is the lack of information with regards to the admission
categories. To overcome this handicap and to make the distinction, Newbold employed an
estimation method that assumed all persons from a particular point of origin were either
immigrants or refugees. This information was then used to extract time series data from the
1990 PUMS relating to those populations. Newbold applied basic statistical techniques to the
temporal data in order to identify and document the similarities and differences in the geographic
distribution and personal attributes between the immigrant and refugee groups he identified in
the PUMS.

Newbold‘s results revealed that the geographic distribution of the foreign-born
population reflected a combined effect of gateway city roles and government resettlement policy.
The gateway city role refers to the notion that some cities have been and continue to be a magnet

86

for immigrants attracting family preferences, immediate relatives, employment, and refugee
arrivals; this is owing to the historical role gateway cities played within the immigration system
supporting large immigrant communities (Newbold 2002). The government resettlement policy
refers to the refugee resettlement policy of dispersion. In Newbold‘s analysis, the policy was
found to be effective because the results indicated that the initial destination of refugee
populations was more dispersed geographically than that of the immigrants. This, however,
changed over time because Newbold‘s results also indicated that their distribution eventually
became increasing concentrated in gateway states and metropolitan areas despite the government
policy of dispersion.

One of the earlier studies concerning the locational determinants for immigrant
settlement was conducted by Ann Bartel (1989). Utilizing Public Use Microdata (PUMS) data
from the 1980 Census of Population and data calculated from Public Assistance Statistics, Bartel
applied a multinomial logit regression to six socioeconomic predictor variables using select
immigrant groups as the dependent variable. She found that the percentage of co-ethnics already
residing in the area had the strongest effect on location choice for immigrants at the SMSA level.
She also found that the more educated individuals within each ethnic group had the highest
mobility rate. The issue with Bartel‘s study, as noted by Newbold (2002), was that the foreignborn were studied in aggregate, thus there was no distinction between immigrants and refugees.

F.H. Buckley (1996) conducted a similar analysis, but at the state level, using ordinary
least squares regression on legal permanent residence data provided by the Immigration and
Naturalization Service, and 1985 through 1994 data from the Statistical Abstract of the United

87

States. Buckley‘s goal was to identify factors attracting immigrants to their immigration state.
The significance of this study was that Buckley separated the immigrant into 4 categories to
include a refugee and asylees category. Buckley‘s findings revealed that his welfare variable –
Aid for Families with Dependent Children (AFDC) – had its strongest positive and only
significant reading with the refugee category indicating that this group is more prone to live in
states with more generous welfare benefits. Buckley cautioned about the findings because of
issues with multicollinearity – one of his predictors was ―correlated with other immigration
predictors‖ (p.88). According to Larose (2006), multicollinearity leads to solution space
instability whereby a set of predictors can result in a regression that is significant overall, even
when none of the individual variables are significant. The use of many predictor variables can
complicate the interpretation of the analysis and violates the principle of parsimony (Larose
2006).

Zavodny (1999) also studied, at the state level, the determinants of immigrant locational
choices for the years 1989 through 1994 using INS and ORR data. The statistical tool for this
study was the linear regression using five predictor variables against six categories of
immigrants. With respect to refugees, the author‘s findings revealed that the welfare and the
total population variables had the strongest positive effect and unemployment had the strongest
negative effect on their location. This confirms Buckley‘s (1996) findings of positive association
with welfare benefits and negative association with unemployment for refugees. Foreign-born
population share had a positive effect, but its strength was nominal compared to welfare and total
population. In this analysis, along with Buckley (1996) and Bartel‘s (1989), prior presence of
foreign-born appears to be the strongest predictor of the traditional immigrant‘s location choice,

88

but it has no real bearing on the location of refugees.

The results of these analyses set

immigrants and refugees apart in their locational choice; however, the results could simply be a
matter of placement for the refugees – while traditional immigrants choose their place of
settlement, refugees have little choice initially.

As mentioned earlier, refugees are placed

according to a set of established criteria aimed at facilitating self-sufficiency like low
unemployment and above average welfare benefits. Thus, the application of welfare magnet
theories may or may not be applicable to refugees – the application depends on the act of
secondary and subsequent migrations.

One of the more comprehensive studies of locational determinants was conducted by
Gurak and Kritz (2000). In an attempt to determine the sources of differentiation between
immigrant/immigrant and immigrant/native-born interstate migration, the authors not only
divided up the immigrant category into twenty-four country of origin groups, they also divided
the variables up into three sets of factors. The variable sets included factors pertaining to human
capital, social capital, and economic conditions of the state of residence. The results from the
logistic regression models revealed that human capital factors were the most important sources
of differentiation in immigrant and native-born internal migration. The study also found that,
while human capital variables provided the greatest explanation for differences in internal
migration, variables associated with the social capital of the groups and the economic conditions
of the state also strongly influenced interstate migration (Gurak and Kritz 2000).

89

Studies Specific to Refugees
While the research mentioned above sometimes involved refugees, the following research
dealt specifically with refugees. Desbarats (1985) was among the first of the researchers to
question the futility of government policies dictating the dispersion of refugees – an issue that
was also brought up by the study done later by Newbold (2002). The questioning came in
response to the increased flow of Southeast Asian refugees who eventually became highly
concentrated in only a few of the Western states. The main objective of Desbarats‘s (1985)
research was to analyze the economic, social, and environmental factors that influenced the
evolution of the Southeast Asian distribution over time. In order to do this, she employed a
backwards stepwise regression. In line with the discussion this dissertation incorporated about
welfare magnets, Desbarats assumed that the refugee‘s locational choices would be a reflection
of the availability and level of welfare assistance. It was also assumed that socio-cultural factors
would be significant in explaining the distribution of the Southeast Asians. Desbarats did not
mention ethnic-origins specifically in her study, but she did rely on some of its propositions to
make assumptions about the expected behavior of her target group based on their cultural
heritage. The assumptions referred to the cohesive nature of refugees ―originating from cultures
marked by tightly knit families whose members are bound by a traditional web of mutual
obligations and aid‖ (p.530).

Desbarats‘s results indicated that the initial placement pattern of refugees showed the
least amount of concentration, but when the refugees were able to assert their own self
determination through secondary migration, their patterns became more concentrated in specific
locations. These locations were found to offer social, climatic, and economic advantages for the

90

refugees (p.535). This move toward a more concentrated residential pattern is what Newbold
(2002) confirmed with his study. According to Desbarats (1985), mild climate states tended to
attract secondary migrants while states with generous assistance programs tended to attract both
primary and secondary migrants. Her study contributed enormously to the geographic study of
refugee distributions in the United States by confirming that the initial stages of ethnic clustering
are crucial in ―polarizing‖ subsequent distribution patterns (p.536). Secondly, her study also
contributed by illustrating the difficulty of slowing the growth of ethnic clusters ―once they have
reached a critical mass‖ (p.536). The final contribution came in the form of advice suggesting
that government policies of refugee dispersion might be more effective if the needs of the
refugees were given more consideration with respect to their spatial placement (Desbarats 1985).

Another such study dealing specifically with refugees was conducted by Simich (2003).
Simich examined the role of Canadian resettlement bureaucracy in ―determining family
connections and shaping the way refugees negotiate social and spatial boundaries during
resettlement‖ (p.576). The policy example given pertained to the political policy that dictates the
dispersal of refugees across Canada – similar to the United States policy. Simich stressed that
such a policy imposed constraints on the refugee‘s ―ability to reconstitute the supportive social
relationships they need for successful settlement. In response to neglect of these needs, refugees
engage in repeated migration‖ (p.576). Simply put, secondary migration is the result of the
refugee‘s application of self-determination as they attempt to locate and reconstitute social
networks as a source of support in spite of ―official destination planning‖ (p.577). Simich
confirmed this notion when her semi-structured interviews and interpretive analysis revealed that
secondary migrations to Ontario took place ―not only for jobs, but more significantly, for social

91

support – by reconnecting with friends and family within the critical first year of settlement‖
(p.587). Simich also found that the decision of the refugees to migrate was made as members of
an ―extended social network‖ – not as individuals (p.582). Simich ultimately stressed that
support from the government is a crucial component in the assistance received by the refugees,
but this support alone cannot meet all of the refugee‘s needs.

Refugees tend to ―maximize their opportunities for social support‖ in a manner that is
specific to their needs with the secondary migration (Simich 2003, 582). Reconstitution of the
family and support network assures an ―emotional and material stability‖ that will last
throughout the resettlement process and longer (p.581). If secondary migration is what it takes
for refugees to become stable and self-sufficient in their resettlement, then a valid question
would be – why would officials responsible for resettling the refugees not resettle them in a
manner that is more accommodating to the refugee‘s needs in the first place? Simich found that
the answer to this involves fault in both parties – the refugee‘s themselves and the officials
responsible for resettlement. According to Simich, the majority of the secondary migrant in her
study ―had clearly expressed destination requests to join family and friends already in Canada,‖
but these requests were not given serious consideration in the final destining decision by
resettlement officials (p.583). In this instance, the fault lies with resettlement officials for
ignoring the request in the attempt to serve political interests by adhering to political policies of
refugee dispersion.

With respect to the refugees, the fault is found when they conceal

information during the application process. According to Simich, Canada has an official family
reunification policy, but many refugees are cautious about revealing family connections – some
believe that ―revealing family support could detract from government financial support or mean

92

that they will not be accepted as refugees at all‖ (p.583). This deception usually results in the
refugee being resettled in a location geographically distant from their social support network.
This displacement of the refugee from his or her support network thus feeds the secondary
migration phenomena.

Research conducted by Hardwick and Meacham (2005) also dealt specifically with
refugees. According to Hardwick and Meacham, in the past decade geographic studies of
refugee issues have emerged as a salient topic of inquiry; they made a contribution by analyzing
the migration experiences and settlement patterns of refugees in the Pacific Northwest. More
and more of the most recent refugee arrivals have settled into smaller towns and cities located in
unexpected places that were long dominated by native-born European populations creating
tension (Hardwick and Meacham 2005). Hardwick and Meacham noted in their analysis of
Portland, Oregon that the large influx of refugees was one of the primary reasons for the
increased diversity seen in that area. The authors believe that an understanding of refugee
settlement patterns is vital because local tensions and ensuing political decisions – which could
affect the lives of not only the foreign-born, but also the native-born populations – are sometimes
a response to the changing residential concentration of the refugees.

With this in mind,

Hardwick and Meacham asserted that a thorough understanding of the refugee‘s residential
pattern requires the understanding of their migration paths, arrival times, cultural and economic
niches, racialization, and the potential subordination by the majority residents (p.540).

Hardwick and Meacham used a multi-method approach integrating both qualitative and
quantitative methods to study the largest refugee groups in Portland – this included refugees

93

from Vietnam and the former USSR. A cartographic analysis of the settlement patterns was
accomplished using census data, refugee resettlement file data, and data from the Immigration
and Naturalization Service. The spatial analysis was fortified by qualitative information collected
through interviews, focus groups, and participant observation – the goal was to help explain the
settlement decision process. Hardwick and Meacham found that a set of ―overlapping and
complimentary political, social, and economic networks influence refugee residential patterns‖
(p.554). In addition, they also found that, regardless of the refugee‘s history, the groups are
initially influenced by the network of support provided by the VOLAGS. After this initial
influence, the various refugee groups were then influenced by different aspects stemming from
their ethnic origins.

Studies Specific to the Somali
There have been studies of the Somali population as evident from the first couple of
paragraphs in this chapter, but not many in the geographical context in which this research needs.
Danso (2001) recognized the limited geographical literature and the predicament motivated his
analysis of Somali refugees in Canada in order to assess their resettlement experiences and
needs. Based on quantitative and qualitative methods, his results suggested that the Somalis
encountered numerous difficulties during the initial stages of resettlement and that information
about resettlement assistance came primarily from their ―personal network of friends, family, and
compatriots‖ (p.3). Danso used a multiple sampling strategy to extract information from the
Somali community. To construct a sample frame, listed subscribers to the Bell Canada telephone
directory who had ―distinctively ‗Ethiopian‘ or ‗Somali‘ last names were randomly picked‖
(p.5). In addition, the snowball sampling method was used to compile subjects for the study.

94

Three sets of survey instruments were used to collect data for this study – one was a
questionnaire, and the other two were semi-structured interviews. The questionnaires contained
open-ended questions that were structured to extract information on pre-migration hopes and
expectations, settlement difficulties, ethnic community organization, linguistic adaptation, labor
market experiences, and residential geographies (p.5).

Danso‘s study revealed the initial

difficulties faced by Somalis to be: affordable housing; finding a job; job insecurity; general
racism; language barriers; insufficient income; and family reunion. These difficulties were
found to be contributors to the high rate of poverty found among the Somalis in Canada.

Danso also made a critical distinction within the Somali population in his study – those
that received government sponsorship, and those that did not.

The Somalis that received

government sponsorship were more informed of resettlement assistance than those who received
other forms of sponsorship. Danso concluded that, constrained access to resettlement assistance
forces Somali immigrants ―to look ‗inward‘ and rely heavily on ethnic networks for information
on employment, housing, education, and language training programs‖ making social networks
invaluable to resettlement (p.12).

Gaps in the Literature
The problem with the studies of the Somali is that they are not locational studies. Studies
of the Somali typically focus on immigration, assimilation, or racial identity as seen with Darboe
(2003) and Kusow‘s (2006) research.

Danso (2001) initiates a geographical aspect in his

research by focusing on issues at the Somali resettlement locations and asserting the value of
social networks at those locations. None of these, however, address matters concerning the

95

locational or migration determinants of the Somali population. There is a gap in the literature
pertaining to the spatial aspects of the Somali – literature that answers not only where they are at,
but also how they got there, and what factors attracted them to that place. The problem with the
locational studies, on the other hand, is that they do not specifically address the Somali.

In

many instances, locational studies either look at immigrants in aggregate thus neglecting the
different categories such as the refugees, or look at refugees in aggregate thus neglecting the
different groups. Separating immigrants into their respective category of entrance for research
was an issue for Newbold (2002) because the different categories in which the immigrants enter
encompass populations with different motivations and different levels of human and social
capital. In addition, analyzing the different categories and groups is relevant because the manner
and timing in which the foreign-born adjust and integrate into American society varies from one
group to the next. There are additional problems with locational studies as they apply to this
research.

First of all, there is the issue of scale – according to Pigozzi (2004), scale is of

fundamental importance because different patterns and causal forces operate at different scales.
So, with respect to the refugees, results seen at the state level might be entirely different from
what may be found at a more local level such as the county, city, or tract level. Secondly, the
statistical tool used and the correlating nature of the data increases the probability of
multicollinearity. Therefore, the significant coefficients found in previous studies might not be
so significant. For these reasons, I employ an analysis that takes place at the county level and
utilizes a statistical tool – principal components regression – that minimizes the effect of
multicollinearity thereby allowing the retention of more variables to better describe the
phenomenon under study.

96

To fill the gap in the literature pertaining to some of the spatial aspects of the Somali
population residing in the United States, this research examines the following research questions;
(1) to determine if Somali migrations are related to clan more so than other socio-demographic
variables, (2) to determine if clans are establishing areas of dominance which could affect the
flow of migration. For example, would members of the Somali who still adhere to clan traditions
avoid certain cities if they know that members of a rival clan make up the majority of the Somali
population in that city? The specific research questions are as follows:

(3) To determine if the cities where the Somali are initially resettled diversified with respect
to clan-families.
(4) To determine if the Somali have dispersed from their original sites of resettlement.
(5) To determine which socio-demographic factors are associated with the distribution of
Somalis.
(6) To determine socio-demographic factors associated with Somali migration within the
United States.
(7) To determine which cities are most associated with specific Somali clans and the factors
related to that association.

97

CHAPTER 5: DATA AND METHODS

To answer the research questions, a combination of both quantitative and qualitative data
was used. The quantitative data was secured from a variety of government database sources and
the qualitative data was derived from surveys. The data and methods are discussed in steps that
are aligned with the research questions.

Step 1: Determination of Clan-Family in Cities of Initial Resettlement
I am seeking to determine if the cities in which the Somali are resettled diversified with
respect to the clan-families that are resettled through the ORR within the United States. With
respect to this research question, my hypothesis is that the cities, in which the Somali are initially
resettled, on average, are highly diversified.

The data for this step came from a special

tabulation from the ORR‘s Worldwide Refugee Admissions Processing System (WRAPS).
WRAPS is a database of information pertaining to the refugees resettled here in the United
States. The special tabulation obtained contains information for fiscal years 1990 thru 2009
relating to the refugee‘s country of origin, year of admission, ethnicity (clan-family affiliation),
state of resettlement, city of resettlement, county of resettlement, and count. I used the ethnicity
field to isolate the Somali by city and by clan-family and a count was totaled for each. A
minimum population threshold of 10 Somali was applied to the dataset to avoid incorporating
communities into the analysis that were likely to immediately lose their resettled population.
During an interview with a counselor from the Community Refugee and Immigration Services in
Ohio, it was revealed that the majority of the time when the Somali are resettled in communities
with less than 10 to 15 Somalis (which is sometimes equivalent to one or two families) they

98

embark on secondary migrations right away – sometimes even a couple of days after arriving in
the community (Staff 2010). A diversity value was assigned to each city using Simpson‘s Index
of Diversity. In ecology, the Simpson Index of Diversity is used to quantify the biodiversity of a
habitat. The primary strength of the index is that it takes into account the aspects of richness and
evenness. Richness refers to the number of species found in a given sample and evenness refers
to the abundance of the different species making up the richness of the habitat. The index for
each city was averaged to test my hypothesis. The index values were also mapped in ArcGIS to
identify any emerging spatial patterns from the values; for example, are there specific regions or
states in which the Somalis are sent to more or less diverse? The Simpson‘s Index of Diversity
(D) is denoted by the following formula (DeJong 1975):

S

D 1 

 n (n  1)
i 1

i

i

N ( N  1)

For the purpose of this research, the Simpson‘s Index of Diversity is used to quantify Somali
clan-family diversity in resettlement cities where S equals the number of clan-family groupings,
ni equals the number of persons belonging to ith clan and N equals the total number of persons in
the sample. Determination of the degree of diversity (high, low, or moderate) was established
using the Jenks Natural Breaks Classification method to determine the best categorization of the
values.

At the most basic level, indices of diversity are applied to determine the variation in
categorical data.

Thus, it is used here to determine the variation in the resettled Somali

99

population. Diversity indices can be used to assess the diversity of any population in which each
member belongs to a unique category. Resultant values range from 0.0 to 1.0 with increasing
diversity as the value ascends. There are many measures of diversity, but the three most
commonly used are the Simpson Index, the total number of species, and Shannon‘s entropy (Hill
1973). Most of the proposed diversity indices – while theoretically useful – have a degree of
bias with regard to sample size, especially when the sample size is small. This issue with sample
bias prompted the widespread use of the Simpson‘s Index because it has been shown to yield
unbiased sample estimates and its sampling variance is known (Smith and Grassle 1977). One of
the major criticisms of the Simpson‘s Index is that it tends to be heavily dependent on the
dominant group within the habitat. The index used in this research represents the probability that
two Somalis picked at random will be members of different clan-families.

The purpose of this step is to establish that there is a high degree of imposed diversity
among the clan-families in their initial resettlement communities. According to Simich (2003), a
survey of secondary migrants in Canada revealed that refugees felt that they should have been
given a choice about their destination, and it would have been more beneficial if the officials
would have respected their ―wishes to join friends or family for reasons of psychological security
and the ability to fill practical needs. If these wishes are not respected, refugees almost certainly
will cross arbitrarily imposed boundaries‖ (p.589). This scenario is pertinent to the Somali
refugees. The Somali have typically been portrayed as coming from an ethnically homogenous
society sharing one religion, speaking one language, and practicing one tradition and culture
(Brons 2001). Lewis (2004), however, notes that their culture of kinship is divisive as evident by
the continuing clan conflict. The way kinships ―are formulated in Somali ideology, have many

100

characteristics of race‖ (p.490). If the Somali are being resettled without location choice, and
their resettlement is based on a Somali national identity without regards to clan affiliation, then
the refugees resettled in the United States could be exhibiting the same sort of drive as the
Somali in Canada. It is possible that they are emigrating from areas where Somali are present,
but the particular Somali are from rival clans. It is also possible that the Somali are taking on
secondary and subsequent migrations not simply to reunite with fellow Somali, but to reunite
with fellow clansmen while at the same time, fleeing rival clansmen.

Step 2: Determination of Somali Dispersion
In this second step, my objective is to determine if the Somali are dispersing (i.e. the state
of being dispersed as in a localized diaspora) out from their sites of original resettlement. The
abundance of research on refugees tends to support the notion that they embark on secondary
migrations to established ethnic enclaves, therefore, becoming more clustered. I believe the
opposite is true of the Somali, thus my hypothesis is - the Somali are dispersing from their
original sites of resettlement therefore becoming more diffused throughout the United States.
Data for this step was secured from the 2000 Census of Population and the WRAPS. The
statistical tool used in this step was the spatial index of relative clustering (RCL). Clustering
indices measure the degree to which racial and ethnic groups live disproportionately in
contiguous spatial units or cluster in space.

According to Massey and Denton (1988), the RCL compares the average distance
between the minority (X) group members with the average distance between the majority (Y)
group members. The resulting value is 0 when X and Y display the same amount of spatial

101

clustering; the value is positive when X is more clustered than Y and negative when X is less
clustered than Y. The relative clustering index is identified by the following formula:

RCL  Pxx / Pyy -1 where
n

n

Pxx  
i 1 j 1

xi x j cij
X2

and
n

n

Pyy  
i 1 j 1

xi y j cij
Y2

This analysis takes place at the county level, so the WRAPS data at the county level is
used. First, I isolated all of the cases from the database that indicate Somalia as the country of
origin for the years 1990 thru 1999. The count for these cases is summed by county and a new
field is created with the Somali count for each county. This new field containing the Somali
count at the county level will be labeled ‗MINOR‘ for use in the RCL formula. It was vital that
the MINOR field be created at the county level so that it would match up with the 2000 Census
county level files. The particular table (QT-P15) of interest from the 2000 Census concerning
the Somali was downloaded in comma separated value format from the American Factfinder
website28 at the county level for each state. Combining the states resulted in a Somali count for
each U.S. County and this comprised the ‗MAJOR‘ field for the RCL. These fields were then
joined to a U.S. county level shapefile via the county‘s fips code. With this shapefile – named
cluster – all the information was present to conduct the relative clustering analysis using the
Segregation Analyzer, a C# Net application for GIS. The MINOR field served as the X and the
MAJOR field served as the Y. What I intend to establish is that the Somali are dispersing from
their initial sites of resettlement by establishing a difference between where they are (MAJOR

28

American Factfinder website: http://factfinder.census.gov/home/saff/main.html?_lang=en
102

field from the census) and where they were actually sent (MINOR field from the WRAPS). I am
looking for a resulting number that is positive which would indicate that the MINOR group is
more clustered than the MAJOR group.

Step 3: Identification of Somali Locational Determinants
The objective of this step is to identify the order of importance for the factors assumed to
be associated with the location of Somalis.

This step deals with social, economic, and

demographic variables, so the data for this step was extracted from various databases. Table 5.1
is a list of the variables used, their aspect, and their description. Sources for the variables are
provided in Table A.3 of the appendix. The aspect category is a general grouping of the
variables that I believe – from reading archive data - carries the most weight in determining
where the Somali are located. I have identified eleven aspects with corresponding variables.

Given the number of explanatory variables and their nature (a high degree of correlation),
the most appropriate statistical tool for determining the association of these variables is a
Principal Components Regression (PCR). According to Cadwallader (1992), the explanatory
variables in such models are often interrelated, thus posing problems of multicollinearity in
estimating the regression coefficients by ordinary least-squares analysis (p.70). Consequently,
Riddell (1970) has suggested that the set of explanatory variables might first be subjected to
principal components analysis (PCA), thus generating a smaller number of orthogonal
components that reflect the underlying similarities of the original variables. These components,
as measured by the component scores, are then used in the regression equation, permitting a
more realistic interpretation of the coefficients.

103

Table 5.1: Variables used for step 3 - locational determinants.
Aspect

Label
wage
1. Employment jobs
unemp
2. Public
housing

3. Public
Assistance
4. Crime
5. Schools

6. Health

7. Climate

subunt

Number of open units under contract for federal subsidy

wait
utilall
fspart
fsben
tanfben
crm
doutr
hosp
ins
phys
mort

Average months on waiting list among admissions
Average utility allowance among households who have it
FSP Participation Per Poor Person
FSP Benefits Per Poor Person
TANF Benefits Per Poor Person
Crime per 1000 persons
Drop-out rate for teens 16-19
Number of hospitals
Percent of persons insured
Number of physicians
Infant mortality rate
Number of climate events causing more than $50,000 in
damages.
Total population
Percent of the population white
Percent of the population black
Percent of the population foreign-born entering between 1990
and 2000
median age of the population
Percent of the population in poverty
Population density (square mile)
Percent of households: family
Vacancy rate for rental units
Percent of households: 3 + bedrooms
Median rent asked
Median real estate taxes
Percent of population voting Democrat
Muslim adherence rate

clmt
pop00
pwhite
pblack

8. Population

9. Households

10. Political
11. Religion

Description
Average wage per job for the county
The estimated per capita number of jobs for the county
The unemployment rate for the county

pfrbrn
mnage
ppov
pdense
fmlyhh
vac
3bed
mrent
mreal
party
muslad

104

PCR is a technique used to handle the problem of multicollinearity and produce stable
and meaningful estimates for regression coefficients (Fekedulegn, et al. 2002); it is a method that
combines linear regression with a PCA. This method transforms a set of correlated variables into
a set of uncorrelated principal components (Liu, et al. 2003).

The components are linear

combinations of the original correlated variables; therefore, the components can be utilized as
explanatory variables in a regression model to obtain parameter estimates, which can then be
used to reconstitute the regression coefficients of the original explanatory variables that are now
corrected in such a way that minimizes the multicollinearity (O‘Brian, Lloyd and Kaneene
1995). The variables in Table 5.1 serve as the independent or predictor variables. The dependent
variable was created from the Somali count from the MAJOR field from step 2. The MAJOR
field was divided by the total number of Somalis from the census and multiplied by 100 to create
a field that is the percent of the Somali population in each county (SOMCEN). The statistical
program used was SYSTAT and the principal component analysis (PCA) – the initial portion of
the PCR – was conducted using Varimax rotation. The random variable method was used to
identify the correct number of dimensions to keep. The random variable (RAND) was used to
illuminate the meaningful dimensions in the set – the dimension in which the RAND variable
had a large loading (loading > 0.6) and all subsequent dimensions were considered meaningless.
The analysis was then run again based on the appropriate number of dimensions and the scores
were saved. The scores were then used in a regression with the dependent variable (SOMCEN).
The standardized coefficients from this model were used in a matrix multiplication process with
the PCA‘s coefficients for standardized factor scores to reconstitute the regression coefficient.
The formula for this action was:

m

R1mm FppC1  m B1

coefficients for standardized factor scores;

p

105

C1

where

m

R 1mm Fp

represented the

represented the model‘s standardized

coefficients; and

m

B1

represented the reconstituted regression coefficients. The generalized

regression equation for this model was y  0  1x1  2 x2  ... k xk   where y represented
the percent of the Somali population in each county, x1, x 2,...x k represented the independent



variables, and 0,1,2,...k represented the regression coefficients to be estimated, and 
represented the error term.






Step 4: Identify Somali Migration Determinants
The objective of this step is to identify and order the socio-demographic variables that
contribute the most to Somali migration. The variables and data (Table 5.2) are similar to those
in step 3, in addition, I added the clan diversity aspect and clan diversity variable from step 1.
The source of the variables is provided in Table A.4 of the appendix. My underlying theme is
that migration is driven by clan motivation; therefore I believe that in these movements, the
places sought out would be places with low diversity indicating a low number of clans or no
clans. The method of analysis and statistical package for use is the same as in step 3 – PCR
analyzed with SYSTAT; however, the source of the dependent variable is different. In this step,
the data for the dependent variable comes from Public Use Microdata Sample files (PUMS) at
the Public Use Microdata Area (PUMA) level. PUMS files are a sample of individual records
from the census long form from the decennial census. PUMS files contain information on
housing units and their occupants that allow users to create tabulations, which could answer a
wide range of social and economic questions about a specific population.

106

Table 5.2: Variables used for step 4 - migration determinants.
Aspect
1. Employment

2. Public
housing
3. Public
Assistance
4. Crime
5. Schools

6. Health

7. Climate

Label
wage
jobs
unemp

Description
Average wage per job for the county
The estimated per capita number of jobs for the county
The unemployment rate for the county

subunt

Number of open units under contract for federal subsidy

wait
utilall
fspart
fsben
tanfben
crm
doutr
hosp
ins
phys
mort

Average months on waiting list among admissions
Average utility allowance among households who have it
FSP Participation Per Poor Person
FSP Benefits Per Poor Person
TANF Benefits Per Poor Person
Crime per 1000 persons
Drop-out rate for teens 16-19
Number of hospitals
Percent of persons insured
Number of physicians
Infant mortality rate
Number of climate events causing more than $50,000 in
damages.
Total population
Percent of the Somali population
Percent of the population white
Percent of the population black
Percent of the population foreign-born entering between 1990
and 2000
median age of the population
Percent of the population in poverty
Population density (square mile)
Percent of households: family
Vacancy rate for rental units
Percent of households: 3 + bedrooms
Median rent asked
Median real estate taxes
Percent of population voting Democrat
Muslim adherence rate
Clan diversity calculated in step 1

clmt
pop00
somcen
pwhite
pblack

8. Population

9. Households

10. Political
11. Religion
12. Clan

pfrbrn
mnage
ppov
pdense
fmlyhh
vac
3bed
mrent
mreal
party
muslad
cland

107

For this research, I used the 5-percent files - this means the data was drawn from a five
percent sample of the population. These samples entail a minimum population threshold for
reasons of confidentiality; the threshold for this data is set at one hundred thousand people and
the size of the population corresponds with the geographic boundaries of the PUMA. PUMAs
follow along the boundaries of counties and rarely cross state or metropolitan boundary lines.
These geographical units do not have formal names; they are identified by a five-digit number
and are a composite of census tracts, places, or counties. The majority of PUMAs are countysized or less; however, in some cases, PUMAs are comprised of several counties – this occurs in
areas where the county population does not exceed the minimum population threshold.

PUMS contains data pertaining to migration; the database can be used to compile a count
of the persons in a tabulated census area who indicated that they lived in different tabulated
census areas 5 years prior to the date the census was taken and where that census area was
located. By creating a matrix for each state‘s PUMAs consisting of the PUMA of residence in
2000 and the PUMA of residence prior to 1995, the state files can be merged to create a master
matrix of all the PUMAs in the United States, their population, and the origins of their migrant
population. This task is accomplished using a DVD purchased from the U.S. Census Bureau of
the Census 2000 Public Use Microdata Sample Census of Population and Housing. The PUMS
2000 DVD

29

contains a copy of Beyond 20/20, a professional browser, which allows for the

creation of cross tabulated extracts of the resident and migration PUMAs by racial or ethnic
group for each state. This DVD was used to create extracts for all those who indicated Somali

29

The files in the PUMS DVD contains the individual weights for each case (persons / housing
unit), which when applied expands the sample to the relevant totals. Thus applying weights to
the PUMS data from this DVD is unnecessary as opposed to using raw PUMS data.
108

ancestry that are migrants. This data was extracted at the PUMA level which, as explained, is of
a different scale than the county level independent variables.

To adequately use the extracted data, I used ArcGIS to join the extracted data to the
Census PUMA shapefile via the five-digit identifier. I then used the U.S. counties shapefile to
parse the PUMA shapefile. This resulted in a file that allowed for the identification of PUMAs
that are equal to or less than a county. The data in the PUMAs equal to a county were left
unaltered and data in each of the PUMAs less than a county were aggregated to the county level.
This resulted in a file of Somali migrants at the county level. The cases in which the PUMAs
were greater than a county were deleted from the file because of the problems associated with
disaggregating data. The newly created file with the county level Somali migrant counts were
then joined with the file of county level independent variables via the county identifiers. Once all
the data was imported into SYSTAT, the Somali count was transformed into the percent of
Somali population similar to what was done in step 3. The PCR was then performed to identify
the variables that contributed the most to Somali migration. Not only did the PCR identify those
variables, it also identified the meaningful dimensions found within the data therefore
illuminating aspects that I may not have conceived.

Step 5: Determination of Cities Associated with Specific Clans
The objective of this step was to determine if there are cities where certain clans
dominate. To answer this question and to paint a broader picture of the Somali secondary
migrations, I drew upon qualitative data derived from surveys, interviews, and participant
observation. I hypothesized that with the increased dispersion of Somalis within the United

109

States, some cities / towns are becoming associated with specific clans. With the available data,
I identified where Somalis were initially resettled by clan-family; I also identified where Somalis
migrate; but, I could not identify where Somali clan-families migrate.

To rectify this

shortcoming, I relied on surveys designed to elicit information about Somali clan-family
migration and settlement. The first survey pertained to organizations and agencies that provide
services to Somalis. I obtained a list of 54 Somali organizations spanning 20 states, which
contains a contact name, telephone number, and email address for each of the organizations.
Additionally, I obtained a contact list for over 200 voluntary agencies spanning 44 states and 151
cities. The surveys for these organizations / agencies were designed to extract information
pertaining to the Somali groups they may serve. The information the survey was designed to
elicit included (but was not limited to) – whether they served multiple clans or was there one
clan in particular that they extend their services to. The survey was also designed to elicit
information on their opinion about the possibilities of a dominant clan existing in their area –
both numerically and politically. If the organizations / agencies are found to serve the Somali,
then the survey would extract information pertaining to which Somali clans they provide services
to. The survey was also designed to determine, – whether or not the organizations / agencies
were dealing with secondary migrants, and if so, were they from a specific clan. The questions
for this survey can be found in Table A.5 of the Appendix. This survey is for distribution in
electronic format utillizing Infopoll web-based survey software. According to Solomon (2001),
web-based surveys offer significant advantages, but there are methodological issues. Of priority
concern was the validity of the data. If one does not ensure precautions, access to the survey
may not be just limited to the subjects but also to those who maliciously or accidentally discover
the survey on the web. In either case, the intended sample could become tainted.

110

The last group I focused on was the Somali secondary migrants in Columbus themselves.
I used a mixture of instruments to obtain information from the Somali community. This included
the use of surveys, unstructured interviews, and participant observation. Questions for the survey
can be found in Table A.6 of the Appendix. Reliance on multiple instruments has become
commonplace in Somali studies because of closed nature of the group. An excellent example of
the use of multiple research instruments can be seen in Bjork‘s study of the Somali practice of
―telling‖ clan affiliation. In the study, Bjork (2007) used a combination of surveys, formal and
informal interviews, and participant observation. The difficulty in using just a single instrument
is best noted by Schaid and Grossman (2007) in their study of Somalis in Barron, Wisconsin.
With respect to surveys, the authors noted that not only was there issues with literacy, there were
also issues with finding gathering points for Somalis to distribute the surveys because of the
residual effects of 9/11.

Schaid and Grossman asserted that, because of raids on Somali

businesses in Minneapolis – in addition to the attempted expatriation of several Somali
community leaders – Somalis all across America went on guard making the survey process
difficult and time consuming. As mentioned earlier, the Somalis are a tight-knit and guarded
community. So prior to designing their survey I first gathered information on the community via
the process of participant observation. I used these observations to guide the manner in which I
created the questions for the surveys and unstructured interviews. The unstructured interviews
were guided by questions from the survey.

Participant Observation
Participant observation is a qualitative method based in ethnographic research, whose
objective is to help researchers learn about the perspectives held by the subject population. This

111

method is distinctive because the researcher learns what life is like for the ―insider‖ while
remaining an ―outsider‖ in the insider‘s own environment (Mack, et al. 2005). The researcher
basically records - in field notes - everything he or she sees as objectively as possible.
Participant observation typically starts with a period of anonymity where the researcher observes
the group under study. In time, the researcher reaches out to a person in the subject population
that the researcher believes could become a valuable asset in gaining information on and contacts
within the community. That person becomes a key informant. I used this approach to procure a
key informant and from the informant, I secured insight into the community as well as contact
information for potential survey recipients and interview subjects.

Snowball Sampling
The nature of the tight-knit community dictated the use of purposive sampling.
Purposive sampling is when the researcher samples with the purpose of selecting specific
predefined groups. A widely used method of purposive sampling is snowball sampling (Wong
2008). Snowball sampling is a term applied to sampling procedures that allow the sample
subjects to provide information about themselves and about other potential subjects (Frank and
Snijders 1994). This method is particularly suited for when the focus of study is on a sensitive
issue or private matters requiring the knowledge of insiders to locate people for study (Biernacki
and Waldorf 1981). The snowball sampling is a non-probability method often employed in field
research where each subject interviewed or surveyed is asked to suggest additional subjects open
to contact (Babbie 1989). This sampling method assumes the subjects are typically members of
social networks, thus when asking them to identify others the interviewer is using the subjects
knowledge to reveal or to identify their social network (Snow, Hutcheson and Prather 1981).

112

This is an efficient method for identifying hidden populations30 or populations that keep low
social visibility. Hidden population refers to a subpopulation of individuals who are unwilling to
disclose themselves (Frank and Snijders 1994). The snowball technique, therefore, allows
resources to be concentrated on interviewing/surveying the most applicable subjects rather than
on screening the entire population to find the target subjects (Snow, Hutcheson and Prather
1981). According to Frank and Snijders ( (1994);
Very few members of a hidden population can usually be found by standard sampling
methods. Often, however, there exists a contact pattern between the members of the
hidden population, which means that they know of know of each other. If these contacts
could be used for finding members of the hidden population, then new estimation
problems arise because of the nonstandard sampling procedure. Snowball sampling is a
way of having initially sampled individuals lead you to other members of the hidden
population, which in turn could lead to further members. (p.53)
Thus a key condition for the successful application of the snowball method is that the
subjects know of others with the traits under instigation and can provide contact information
(Kalton and Anderson 1986). This method was applied because of the secretive nature of the
Somali population when issues concerning clan are brought up by those that are not Somali – i.e.
outsiders. Even though this method is most appropriate, there are some assumed biases due to
the method‘s nonrandom nature. Welch (1975) lists three kinds of bias associated with snowball
sampling: (1) under-sampling of isolated members of the community and oversampling of those
with more extensive contacts. This potentially leads to (2) bias in the socioeconomic class of the
respondents and (3) bias against those persons with views that are unpopular within the
community. Petersen and Valdez (2005) however, insists that the biases can be minimized by:
(1) avoiding institutional references, (2) maintaining high visibility within the target population,

30

Hidden populations refer to subsets of a population whose membership is not easily
distinguished based on existing knowledge or for sampling capabilities (Petersen and Valdez
2005).
113

(3) making frequent social contacts within the target community, and (4) using community
gatekeepers. Biases can also be reduced by using the exponential non-discriminant form of
snowball sampling (Figure 5.1). The bias noted by Welch (1975) typically occurs when using
the linear form of snowball sampling. However, by using the exponential non-discriminant
form, I increased the likelihood of including multiple networks and different social classes within
my sample population.

Figure 5.1: Forms of snowball sampling.

Exponential
Non-Discriminant

Linear

Again, when working with hidden populations, using more conventional sampling
techniques are often inadequate for producing data, leaving researchers to employ other
strategies. Snowball sampling methods are the preferred method when studying groups whose
members are hard to identify. The nature of the method, however, means the samples are neither
random nor representative, thus resulting in selection bias and limitations in validity (Petersen
and Valdez 2005). And, since the samples are non-probability samples, objective weighting
adjustments cannot be employed to compensate for this factor (Kalton and Anderson 1986).

114

With regard to the results;
When research concerns a known population, the choice of sample design is dependent
on individual research goals. Whereas quantitative designs use ‗representative‘ sampling
strategies to make inferences about a whole population, qualitative sampling designs are
non-generalizable, but provide maximum theoretical understanding of social process.
(Biernacki and Waldorf 1981)

Thus, the results cannot be used to make generalization about all Somali secondary migrants, but
it can be used for understanding the secondary migration phenomenon.

Unlike the first surveys, the survey for the Somali was in hard-copy format to minimize
some of the biases that might be apparent with the use of electronic media. The first two surveys
were web-based without overwhelming fear of biases because one of the modes of contact was
by e-mail. This action assumed that the recipients did have access to the Internet. All three
surveys consisted of twenty to twenty-five open and close-ended questions requiring ten to
twenty minutes to complete. The surveys were of enumerative inquiry. This approach involved
the listing or classifying of items by percentages, frequencies, rank order or whatever was useful
in solving the research question (Grbich 2007). The test for survey reliability was accomplished
through the alternate-form method, which involved imbedding the survey with duplicate
questions measuring the same attribute, but worded differently. The validity of the survey was
tested via the face and content validity methods whereby the survey was subjected to review by
both trained and untrained individuals in the field of survey research. Finally, a pilot test of the
survey was conducted to ensure the survey measured what it was supposed to measure. The
survey portion of the research adhered to protocols set forth by the Institutional Review Board to
protect the identity and confidentiality of the research subjects.

115

Columbus, Ohio as Case
The location chosen for the participant observation, surveys, and interviews to take place
was Columbus Ohio. Columbus represented a unique opportunity for a couple of reasons; first
and foremost is the fact that Columbus was recently the destination point for a large Somali
refugee secondary migration – the migrants had no apparent links or prospects in the city.
Second, Columbus was an ideal place for study because of its proximity to this researcher.
Finally, Columbus has the second largest grouping of Somalis in the United States, which
increased the potential pool of survey subjects.

Table 5.3 contains the most recent data from the American Community Survey pertaining
to the demographic and social characteristics of Columbus Ohio. By all accounts, Columbus is
economically thriving and was ranked the eighth best big city in 2006 by Money Magazine
(CNN Money 2011). Columbus has a total population of over 750,000 people – 66.0 percent
white, 25.7 percent are black, and 4.0 percent Asian. Out of this population, 4.6 percent are
Hispanic, and 9.5 percent of the population is foreign-born. Columbus has a fairly young
population, the median age is estimated at 31.2 which is below the median age for the country.
From an educational aspect, over 87.0 percent of the population had a high school diploma or
higher and 32.5 percent had a Bachelor‘s degree or higher.

Median family income and per

capita income are lower than the national average; however, the difference is nominal given the
low cost of living in Columbus. The negative aspect of the city concerns the poverty rate – both
individual and family poverty rates (19.9 and 14.8 respectively) are well above the national
average (13.5 and 9.9 respectively).

116

Table 5.3: ACS demographic estimates - 2005-2009 for Columbus, OH.
Population Characteristics
Total population
White
Black
Asian
Hispanic
Social Characteristics
Avg. family size
Median age (years)
Population 25 years and over
High school graduate or higher
Bachelor's degree or higher
Foreign born
Economic Characteristics (in
2009 inflation-adjusted dollars)
In labor force (16+)
Median family income
Per capita income
Families below poverty level
Individuals below poverty level
Housing Characteristics
Total housing units
Occupied housing units
Owner-occupied housing units
Renter-occupied housing units

Estimate
753,572
497,387
193,688
30,121
34,464
Estimate
3.07
31.2
490,808
(X)
(X)
71,796
Estimate
417,995
52,917
23,472
(X)
(X)
Estimate
367,419
313,416
159,047
154,369

Percent
66.0
25.7
4.0
4.6
Percent
(X)
(X)
65.1
87.0
32.5
9.5
Percent
70.2
(X)
(X)
14.8
19.9
Percent
85.3
50.7
49.3

U.S.
74.5%
12.4%
4.4%
15.10%
U.S.
3.1
36.5
84.6%
27.5%
12.4%
U.S.
65.0%
62,363
27,041
9.9%
13.5%

Margin of
Error
+/-323
+/-2,223
+/-1,897
+/-1,003
+/-1,180
Margin of
Error
+/-0.0
+/-0.2
+/-1,519
(X)
(X)
+/-2,490
Margin of
Error
+/-3,032
+/-799
+/-250
+/-0.7
+/-0.7

Margin of
Error
+/-1,570
88.2%
+/-2,096
66.9%
+/-1,814
33.1%
+/-2,217

U.S.

Source: (FILES: 2005-2009 American Community Survey Ohio / prepared by the
U.S. Census Bureau 2011).

117

On the positive side, over 70.2 percent of the population that is 16 years old and over are in the
labor force. At the beginning of 2011, Columbus had an unemployment rate of 8.3 which was
lower than the rate for the State of Ohio – 10.1 (The Columbus Dispatch 2011).

Columbus represents an ideal resettlement community, and since 1990, the ORR has
resettled over 4917 Somali in the city (Figure 5.2). The Somali resettled in Columbus have come
from diverse groups. Table 5.4 is a compilation of ORR data arranged according to the top 10
groupings from the Somali refugee population resettled in Columbus between 1990 and 2009.
Figure 5.2 shows the Somali resettlement count starting in 1993 because that is the first year the
database officially recorded the Somalis as being resettled in Columbus. Finding the count of the
Somali population resettled in Columbus is a straight forward process, but figuring out the
Somali population that actually lives in Columbus, on the other hand, has been a major issue
because of secondary migrations. Census estimates place the population at over 10,000; Somali
born literature places the count at 45,000; and Somali organizations have eluded to numbers as
high as 80,000 (Community Research Partners 2009). Given such divergent estimates, the
Community Research Partners (CRP) issued a report in which they used the latest census data
and conducted tests on two other estimation methods to derive at a more accurate estimation of
the Somali population in Columbus. According to the report, The American Community Survey
(ACS) put estimates at 7,585 ± 1922, the CRP student-data method placed the population at
13,926 for the year 2007, and the CRP birth-data method placed the 2007 Franklin County
Somali population between 13,311 and 41,338 – Columbus would account for 90 percent of the
births (Community Research Partners 2009).

118

Figure 5.2: Somali ORR resettlement in Columbus by year.
SOMALI RESETTLEMENT IN COLUMBUS 1990 - 2009
1400

NUMBER

1200
1000
800
600
400
200
0

Table 5.4: Top 10 Somali groups resettled in Columbus, OH 1990 - 2009.
Rank
Group
1
Darood
2
Other
3
Tuni
4
Barawan
5
Midgan
6
Benadir
7
Tumale
8
Hawiye
9
Rahanweyn
10
Isaaq
Total for top 10 resettled Somali groups:

Number
1801
1758
266
250
160
129
113
54
36
26
4593

Percent
36.6
35.8
5.4
5.1
3.3
2.6
2.3
1.1
0.7
0.5
93.4

A total of 4917 Somalis were resettled in Columbus, those that did
not identify their ethnic grouping were classified as ―unknown.‖

119

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1993

SOMALI

The report concluded that the most accurate figure would place the Franklin County Somali
population for the year 2007 at 15,000 which translates into a Somali population of 13,500 for
Columbus which is more reasonable, and within the lower bracket of the other estimates.

While the actual population count remains an issue, their spatial distribution in Columbus
is not (Figure 5.3). The vast majority of Somali – even though they can be found residing in
many neighborhoods in Columbus (the larger map at the census tract level) – are fairly
concentrated in the northeast section of the city in the Linden area as revealed by the Hot Spot
Analysis (the second smaller map). Many of the Somali refugees were placed in public housing
projects in this area by resettlement agencies because it was a predominantly lower-class African
American neighborhood. According to Riley (2005), the assumption in placement was that the
two groups would get along based on skin color, however, these assumptions were found to be
wrong as the city faced an escalation in tension (which is still apparent today) and conflict
between the two groups over space and resources. While the tension was initially thought to be
group specific differences, the research conducted by Riley uncovered information pointing to
deeper issues – issues concerning the distribution of resources.

The African American

population felt slighted because of the resources (both federal and local) that have been provided
to the Somali in Columbus (Riley 2005). These resources include local business start-up funds
from local agency/city partnerships. Thus, a large part of the tension revolved around the access
to capital that was made available for the Somali, but not for the African American population.
This tension was fueled by the notion that the vast majority of Somalis are not citizens, yet their
access to financial resources was better than that of citizens.

120

Figure 5.3: Somali residential distribution in Columbus.

121

Limitations of the Data
There are three main limitations of the data. First, there is the issue of time. The census
and PUMS data both are for the census year 2000. The census and PUMS represent a snapshot
of a period in time. At that period in time, only one-third of the Somalis currently residing in the
United States were present for enumeration. Because of this limitation, my analysis cannot
illuminate the conditions that currently affect the Somali migrants with respect to locational
determinants. A second limitation of the data deals with scale. While this analysis is at a level –
to this researcher‘s knowledge that is – lower than any previous research, I faced a possible loss
of pertinent data when converting over from PUMA to county level Somali migrant counts in
step 4. Step 4 dictated the removal of those PUMAs (and the enclosed county counter-parts)
where the PUMA was comprised of more than one county because of issues with disaggregating
data. The PUMAs that were removed represented the extreme rural areas and the impact to the
overall Somali count is likely to be nominal. However, this represents a limitation because I was
unable to use the entire dataset (in step 4) with the deletions, which could affect the predictive
power of the regression model. A third limitation of the data is the lack of specification of
Somali clan/ethnicity in the PUMS. PUMS data lists Somalis in their migration tables, but it
does not list clan affiliation. If clan affiliation were listed, this would have added definitive
strength to my argument that Somali migrations are associated with clan movement. In light of
this limitation, reliance was placed on responses to the surveys to highlight destination points for
the Somali migrants with respect to clan affiliation. This limitation can also be seen in the clan
diversity index, which was derived because there is no specific information in the census about
the Somali clan-families that would enable me to quantify the difference between where clans
were sent with the Worldwide Refugee Admissions Processing System data, and where they

122

moved to with the census data. Therefore, I relied on the magnitude of clan diversity to indicate
clan association with migration when taken in consideration with information extracted from
their ethnic origins.

123

CHAPTER 6: RESEARCH FINDINGS

Determination of Clan-Family Diversity in Cities of Initial Resettlement
There were 179 cities included in this analysis of the clan-family diversity. Results from
the Simpson‘s Index of Diversity (Table 6.1) reveal that the index for the 179 cities, on average,
was 0.467 for the 1990 to 1999 admission period. This indicates a medium level of diversity.
The results represent the probability that two Somalis picked at random will be members of
different clan-families.

Table 6.1: Resettlement city diversity (N=179).
Simpson's Index of
Diversity
Years

Index

1990 - 1999
1990 - 2009

0.467
0.578

Paired Samples T-Test
Std.
Deviation
0.270
0.273

t

df

Sig. (2tailed)

-8.373

178

0.000

Jenks Natural Breaks Category of Diversity.
0-0.154
low level diversity
0.155-0.556
medium level diversity
0.557 - 1.00
high level diversity

This step of the research however is concerned with the more inclusive timeline of data
ranging from the start of the refugee flow in 1990 to the more recent round of admissions in
2009. Results for the 1990 to 2009 admission period revealed a diversity index of 0.578 which is
a high level of diversity. This result lends support to my overall theory that Somali migrations
within the United States are influenced to some degree by clan factors. The diversity in all cities
included in this analysis, on average, saw an increase in clan-family diversity. The Simpson‘s

124

Index of Diversity for the cities for the period ending in 2009 was, on average, 0.026 points
higher than the period ending in 1999.

A Paired Samples T-test successfully revealed a

statistically reliable difference between the diversity results for these two periods. The increased
diversity can be attributed to the increased count of Somalis being admitted after the year 2000.
This can also be attributed to family reunification whereby the refugees previously resettled fill
out affidavits of relationship to have family members join them. If the family extends across
clan-families, the affidavit could serve to increase the number of clan-families as well as their
population count within the cities.

Determination of Somali Dispersion
Table 6.2 shows the results of the relative clustering analysis performed with the
Segregation Analyzer. The result of this analysis is 0.253 which is a positive number that
indicates that the MINOR group is more clustered than the MAJOR group.

TABLE 6.2: Index of relative clustering.
Test
RCL

Group
MINOR
MAJOR

Minor
0.000
0.253

Major
0.253
0.000

The interpretation of the result from this analysis leads to the conclusion that the average
distance between the Somali population based on their initial resettlement locations was less than
the average distance between the Somali population based on their locations for the 2000 Census
count. The knowledge that the Somali population tended to be less clustered in the places they
were found at versus the places they were sent to provides evidence of spatial dispersion from

125

the initial resettlement communities as hypothesized. Using ethnic origins as a foundation, I
based my hypothesis on the historic behavior of the Somali people occupying specific territory in
their home region by clan-family. It is my belief that given the limited amount of control over
where they were sent when resettled in this country, the Somali enacted their own redistribution
reconstituting supporting ties with fellow clansmen thus mimicking the spatial relationships
similar to what was witnessed in Somalia. This would also fit with what Simich (2003) found
with her research as she concluded that the displacement of refugees from their support network
fed the secondary migration phenomenon.

Identification of Somali Locational Determinants
Results for the identification of the locational determinants are found in Tables 6.3, 6.4,
and 6.5. The variables used to identify the locational determinants were first subjected to the
PCA. The PCA portion of the PCR was conducted using the random variable method to identify
the number of meaningful dimensions to extract. This was followed by a varimax rotation.
Varimax rotation is an orthogonal rotation – the result was the extraction of dimensions that were
uncorrelated. The random variable method identified nine meaningful dimensions for rotation –
these orthogonal dimensions combined accounted for 73.6 percent of the total variance found
within the dataset. The nine orthogonal dimensions, along with their corresponding variables
with high loadings (> 0.6), are found in Table 6.3. In this orthogonal analysis, the loadings are
equivalent to the correlation between the observed variables and the dimension.

126

Table 6.3: PCA Locational determinants - dimensions, loadings, and variance (n=1667).
Dimension 1: Level of Residential
Development (10.75%)
Total population
Number of hospitals
Number of physicians
Dimension 3: New Immigrant Magnet
(6.39%)
Percent of the population foreignborn entering between 1990 and
2000
median age of the population

Dimension 2: Welfare Generosity
(7.62%)
0.942 FSP Participation Per Poor Person
0.912
0.934 FSP Benefits Per Poor Person
0.909
0.860
Dimension 4: Family Accommodations
(6.80%)
0.636 Percent of households: family
-0.597

Percent of households: 3 +
bedrooms

-0.881
-0.658

Crime per 1000 persons
0.578
Dimension 5: Regional Economic Effects
Dimension 6: Regional Employment
(11.64%)
Effects (10.40%)
Median real estate taxes
0.747 Percent of persons insured
-0.822
Percent of the population in
Median rent asked
0.741
0.751
poverty
The unemployment rate for the
Vacancy rate for rental units
-0.707
0.722
county
Average wage per job for the county
0.658
Dimension 8: Demographic Effects
Dimension 7: Housing Availability (7.31%)
(8.84%)
Population density (square miles)
0.935 Percent of the population black
0.895
Number of open units under contract
0.927 Infant mortality rate
0.783
for federal subsidy
Percent of the population white
-0.728
Dimension 9: Public Housing Demand
(3.84%)
Average months on waiting list
0.555
among admissions

127

Table 6.3 contains the first part of the results concerning the locations of the Somali. The
dimension that explains the greatest amount of variation in the data (11.6 percent) is Dimension
5 which has been named ―Regional Economic Effects‖ because the variables contained within
this dimension are more sensitive to the effects of the economy. In this dimension, median rent,
median real estate taxes, and average wages are positively correlated with the dimension while
the vacancy rate variable is not. The dimension that accounts for the next largest amount of
variation is Dimension 1 – ―Level of Residential Development‖ – which accounted for 10.75
percent of the variation in the dataset. Dimension 1 contains variables pertaining to population
size and some of the attributes associate with population size like a corresponding number of
hospitals for the population and a corresponding number of physicians for the hospitals. The
next dimension accounting for third largest percentage of the variation is Dimension 6 – 10.40
percent of the variation. Dimension 6 – or ―Regional Employment Effects‖ – contains variables
directly associated with employment – unemployment, insurance, and poverty.

The

unemployment and poverty have a positive correlation to the dimension. These dimensions
where mentioned first because out of the amount of variation found in the dataset, these three
dimensions combined account for close to one-half of that variation. One of the more interesting
dimensions to note is Dimension 8 – ―Racial Demographic Effects.‖ The variables with the
highest loadings in this dimension include the percentage of the population that is black and
white, and the infant mortality rate. What Dimension 8 reveals is the association between race
and structural inequalities. The effects of this type of distribution, where the white and black
populations are inversely related and the black group is dominant, is commonly reflected in the
healthcare system as evident by a corresponding high infant mortality rate. Other dimensions of
special interest are Dimensions 2, 3, and 4. Dimension 1 already infers locations that are highly

128

populated – Dimensions 2, 3, and 4 narrow the locations down to areas that are characterized by
a high level of welfare usage and housing that is not family oriented. The dimensions also reveal
that the locations are characterized by a younger population with a high degree of new immigrant
arrivals and a high degree of criminal activity. Taken collectively, the extracted dimensions and
their corresponding variables tend to describe locations with central city characteristics.

Table 6.4 contains information pertaining to the regression on the PCA scores from our
nine dimensions. While the nine dimensions described earlier are extremely informative about
Somali locations, Dimensions 2, 6, 7, 8, and 9 appear to be not very significant. Dimensions 1,
3, 4, and 5, however, are significant and the effects of multicollinearity in all dimensions have
been neutralized as evident by the tolerance levels. The effect of the dimensions that were found
to be not significant is not readily evident in the ordering of their variable‘s reconstituted
coefficients in Table 6.5. In the table, those variables occupy positions throughout the ordering –
with the exception of the top five.

Table 6.5 contains the reconstituted coefficient results of the PCR. The reconstituted
coefficients are ordered because the objective is not to produce the largest possible coefficient of
determination in reference to the R2, but rather to understand and explain the phenomenon under
analysis (Riddell 1970).

What we see in this table is that the locational determinants of the

Somali appeared to be heavily influenced by variables from Dimension 1 concerning the level of
residential development. The variable having the most influence on the location of the Somalis
is the number of hospitals.

129

Table 6.4: Regression on locational dimension scores.
Dependent Variable
N
Multiple R
Squared Multiple R
Adjusted Squared Multiple
R
Standard Error of Estimate
Effect Coefficient
CONSTANT
Dimension 1
Dimension 2
Dimension 3
Dimension 4
Dimension 5
Dimension 6
Dimension 7
Dimension 8
Dimension 9

0.06
0.13
0.01
0.04
0.07
0.07
-0.03
0.01
0.01
-0.01

SOMCEN
1,667
0.27
0.07
0.07
0.62
Standard
Std.
Tolerance
Error Coefficient
0.02
0.00
.
0.02
0.21
1.00
0.02
0.02
1.00
0.02
0.06
1.00
0.02
0.10
1.00
0.02
0.11
1.00
0.02
-0.04
1.00
0.02
0.01
1.00
0.02
0.01
1.00
0.02
-0.01
1.00

130

t p-value
3.57
8.66
0.83
2.68
4.33
4.77
-1.66
0.47
0.43
-0.35

0.00
0.00
0.41
0.01
0.00
0.00
0.10
0.64
0.67
0.72

Table 6.5: PCR results - locational determinants.
Order Variables:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

Number of hospitals
Total population
Number of physicians
Percent of household - family
Estimated per capita number of jobs for the county
Population density (square mile)
Percent of the population foreign-born entering between 1990
and 2000
Average months on waiting list among admissions
Number of units under contract for federal subsidy and
available for occupancy
Percent of households with 3 + bedrooms
TANF Benefits Per Poor Person
Average wage per job for the county
Muslim adherence rate
Median rent asked
The unemployment rate for the county
Percent of population voting Democrat
Average utility allowance among households who have it
Crime per 1000 persons
median age of the population
Percent of the population white
Number of climate events causing more than $50,000 in
damages.
FSP Benefits Per Poor Person
Percent of the population black
Median real estate taxes
Vacancy rate for rental units
Percent of persons insured
FSP Participation Per Poor Person
Percent of the population in poverty
Infant mortality rate
Drop-out rate for teens 16-19

131

Reconstituted
Coefficients
0.066
0.063
0.047
-0.041
0.039
-0.030
0.027
-0.026
-0.026
-0.025
0.025
0.023
0.022
0.018
-0.018
0.017
-0.016
0.014
-0.013
-0.012
0.012
0.008
0.008
0.007
-0.006
0.006
0.005
-0.004
-0.002
-0.002

This is followed by the total population and the number of physician variables. While these are
Dimension 1 variables referencing residential development, the hospital and physician variables
can also be taken to represent a healthcare dynamic. The healthcare theme for the top variables
can be explained by the refugee category under which the vast majority of Somalis enter. One
of the main concerns of the resettlement program is that the refugees have access to healthcare
facilities because of the documented trauma – both mental and physical – they have endured.
Thus, VOLAGS tend to place emphasis on locations with proper medical facilities in their
resettlement decisions. The healthcare related variables are not the only variables in this table
that can be attributed to the effects of refugee resettlement agendas – nearly all variables in the
top half of this table can be linked to an idealized attribute of a refugee resettlement community.
For example, the next two variables at the top of the list having the strongest impact on Somali
locations are percent of households that are family households and per capita number of jobs.
The main goal of refugee resettlement is the facilitation of self-sufficiency, so communities with
an adequate number of jobs is paramount. Thus the per capita number of jobs is positively
associated with Somali locations; however, family households are not.

This can also be

explained through the resettlement process. Most of the refugees are resettled in close proximity
to the local agency responsible for their well-being – this is typically in central cities where the
percentage of family households are low as opposed to outer lying suburban areas. The top half
of Table 6.5 has other variables referring to housing such as the percent of households with 3+
bedrooms which is also negatively associated with the Somali distribution. This can also be
explained as a result of the resettlement process and the central city location. It would not be
uncommon to resettle the Somali in public housing projects within the city limits as was the case
with the Somalis in Columbus – the vast majority were initially resettled in the Linden projects

132

which at the time was predominantly African American. In many other instances, large numbers
are resettled in low income apartment complexes – as more and more are resettled in these
apartments, the Somali population quickly becomes the resident majority.

Most of these

building are not family oriented because they are mainly efficiencies, and one-or-two bedrooms.

The table is also dominated by variables such as subsidized housing available for
occupancy and subsidized housing waiting lists. The negative association and the central city
explanation of the aforementioned housing variables can be applied here as well. Another
variable occupying a position of high importance is the percent of recently arrived foreign-born.
Again, this can be contributed to the resettlement process of locating the refugees within the city
– not only the Somali, but also refugees from other countries. The next couple of variables, even
though they hold less importance than the variables already mentioned, still have a strong
influence on the location of Somalis. For example, the temporary assistance for needy families
benefit variable – an idealized variable for resettlement communities. This also applies to the
unemployment variable to which there is a negative association. There is also the median rent
asked variable which can be explained by a central city location. One interesting variable found
to have a strong influence as a Somali locational determinant is the Muslim adherence rate
variable. The overwhelming majority of Somali adhere to the Muslim faith, but its placement in
the table suggests that many other factors take precedence.

The second half of the variables on the table, although holding less importance, still tells
a lot about the Somali distribution. These variables include factors such as crime, median age,
poverty, welfare benefits, and the percent of the population black and white. While the Somali

133

distribution is negatively associated with variables like poverty, infant mortality, and drop-out
rates, it is positively associated with several welfare variables and crime. In addition, while the
Somali distribution is positively associated with the percent of the population that is black
(which can be linked to the resettlement practice of resettling Somalis in close proximate to
African Americans), it is negatively associated with the percent of the population that is white –
a variable that held more importance in the table. Many of the variables seem to work in
opposition to one another with respect to the negative or positive influence they have on the
location of the Somali, but this can be explained by the spatial location of Somalis within the
urban environment – more specifically, within central cities. When considering the nature of
some of the variables and the direction of association – this leads me to conclude that the
distribution of the Somali was still heavily influenced by the locations the refugee resettlement
organizations established. This conclusion is logical when considering that the data was from a
period when the Somali could have been still considered as ―new‖ to America. The visible
Somali flow into this country began in the early 1990s, but large scale entry did not occur until
after the year 2000 – all data used in this analysis was for the year 2000. Table 6.5 revealed that
the Somali locations were positively associated with the welfare variables, but negatively
associated with variables such as poverty and unemployment. Their distribution was positively
associated with new foreign-born, blacks, and crime but negatively associated with whites,
family households, and family type housing. With this group being ―new‖ to America, the
contrast in the variables could not be attributed to ethnic enclaves – during this period the Somali
population would have been in the process of enclave building. Thus, a plausible conclusion is
that the determinants of Somali location at this point in time were resettlement community
related. In order to facilitate rapid self-sufficiency, VOLAGS resettle refugees in places where

134

the communities can absorb the influx. These places are typified by low unemployment rates,
low poverty, generous welfare systems, and the availability of jobs, and the availability of
housing. In a spatial context, the VOLAGS typically resettle refugees in central areas that are
characterized by a higher proportion of blacks, a low proportion of whites, a high degree of new
foreign-born, and a younger population. Housing is available, but the amenities are rarely family
oriented in that they are typically one to two bedroom apartments that are, in many instances, not
subsidized by the federal government, but rather by the local or state governments. Given the
characteristics of resettlement communities, we can explain most – if not all – of the variables in
the table affecting the Somali distribution both negatively and positively.

Identification of Somali Migration Determinants
The results for the identification of migration determinants are found in Tables 6.6, 6.7,
and 6.8. In a similar fashion like the last section, Table 6.6 shows the results of nine orthogonal
dimensions extracted during the PCA portion of the PCR and their corresponding variables with
the high loadings (>0.6). These nine dimensions combined accounted for 71.1 percent of the
total variance in the dataset. These are dimensions associated with Somali migration and the
dimension that accounts for the greatest amount of variability in the dataset (12.2 percent) is
Dimension 1 – which was also named ―Level of Residential Development‖ much like in the last
section because many of the variables with high loadings was associated with the residents of the
migration destination. The variables in question include population size, number of hospitals
and physicians. The important variable to note in this first dimension, however, is the ―clan
diversity‖ variable.

135

TABLE 6.6: PCA Migration determinants - dimensions, loadings, and variance (N=1667).
Dimension 1: Level of Residential
Development (12.24%)
Total population
Number of hospitals
Number of physicians
Clan diversity

Dimension 2: Welfare Generosity (7.21%)
0.915 FSP Participation Per Poor Person
0.896 FSP Benefits Per Poor Person
0.851
0.671

Dimension 3: New Immigrant Attraction
(4.09%)
Percent of the population foreign-0.670
born entering between 1990 and
2000

0.913
0.907

Dimension 4: Regional Economic Effects
(10.52%)
Median rent asked
0.736
Median real estate taxes
Average wage per job for the
county
Vacancy rate for rental units

0.724
0.644
-0.707

Dimension 6: Regional Employment Effects
(9.63%)
0.908 Percent of the population in
0.748
poverty

Dimension 5: Housing Availability (6.92%)
Number of open units under contract
for federal subsidy
Population density (square mile)

0.906 The unemployment rate for the
0.692
county
Percent of persons insured
-0.821
Dimension 7: Family Accommodations
Dimension 8: Racial Demographic Effects
(6.48%)
(8.48%)
Percent of households: 3 +
-0.644 Percent of the population black
0.898
bedrooms
Percent of households: family
-0.878 Infant mortality rate
0.786
Percent of the population white
-0.734
Dimension 9: Criminal Propensity (5.49%)
Crime per 1000 persons
Median age of the population

0.647
-0.562

136

This variable was based on the Simpson Index of Diversity calculated at the county level. While
most of the dimensions and variables were similar to that seen in Table 6.3, there were some
differences noted. For example, Dimension 3 – the ―New Immigrants Attraction‖ dimension lost
two of its variables which now formed their own dimension requiring a new label. The crime
and median age variables occupy Dimension 9 which was labeled ―Criminal Propensity.‖
Dimensions 3 and 9 account for the least amount of variation (4.09 and 5.49 percent
respectively) found in the dataset. An additional change includes the variable that remained in
Dimension 3 – percent of the foreign-born recently arrived. This foreign-born variable is no
longer positive – it is now negative in this model. This could be indicative of a change whereby
the Somali migrants are moving away from the contemporary immigrant receiving areas to areas
where the immigrants are more established. The final major change noted occurred with the
―Regional Economic Effects‖ dimension which expanded. The ―Regional Economic Effects‖
theme now occupies Dimension 4 (formerly Dimension 5) and it accounted for the second largest
amount of variation (10.52 percent) in the dataset. Dimension 4 now contains the median rent
and median real estate tax variables – both are positive and both share a relatively equal high
loading. The ―Family Accommodations‖ theme that used to occupy Dimension 4 now occupies
Dimension 7.

Table 6.7 contains the information pertaining to the regression on the PCA scores from
our nine dimensions.

The analysis in this instance was quite different from the analysis

performed on the PCA for the locational determinants. First of all, the R2 is much higher –
though the concern is not with producing the highest coefficient of determination. Secondly,
there were only two dimensions that were found to be not significant – Dimensions 6 (Regional

137

Employment Effects) and 8 (Racial Demographic Effects). The effect of the dimensions that
were found to be not significant is evident in the ordering of their variable‘s reconstituted
coefficients in Table 6.8. In the table, with the exception of one, the variables belonging to the
dimensions that were found to be not significant occupy a place at the bottom of the ordered
reconstituted coefficients. Again, the tolerance levels for all dimensions were impeccable.

Table 6.8 contains the results for the reconstituted coefficients for the Somali migration
determinants. As evident in the table, most of the ―Residential Development‖ variables are not
as influential as they were in the locational model. An ordering of the coefficients indicate that
those variables pertaining to the Somali and the recently arrived foreign-born dominate.
According to the table, Somali migrants are most influenced by destinations with existing Somali
populations. Secondly, the migrants are influenced by destinations characterized by diverse
clan-family mixtures. This result provides evidence that runs in opposition to my overall beliefs
about the behavior that would be displayed by the Somali migrants. Even though I hypothesized
that clan membership heavily influences the migration of the Somali, in this particular instance,
my hypothesis was that the migration would be characterized by movement to areas where the
level of diversity was low. Thus, I expected a negative impact for clan diversity variable.
Again, it must be stated that the data is from a period before the bulk of the present Somali
population arrived – a period of initial enclave building such as in places like Columbus, Ohio
and Minneapolis, Minnesota. This period of enclave building is supported by the results of the
city diversity analysis which showed the increase over time.

138

Table 6.7: Regression on migration dimension scores.
Dependent Variable
N
Multiple R
Squared Multiple R
Adjusted Squared Multiple
R
Standard Error of Estimate
Effect
CONSTANT
Dimension 1
Dimension 2
Dimension 3
Dimension 4
Dimension 5
Dimension 6
Dimension 7
Dimension 8
Dimension 9

Coefficient
0.05
0.26
0.07
-0.27
0.09
-0.11
-0.02
0.12
0.02
-0.11

PSOMMIG
1,667
0.64
0.41
0.41
0.53
Standard
Std.
Tolerance
Error
Coefficient
0.01
0.00
.
0.01
0.39
1.00
0.01
0.09
1.00
0.01
-0.39
1.00
0.01
0.13
1.00
0.01
-0.16
1.00
0.01
-0.03
1.00
0.01
0.18
1.00
0.01
0.03
1.00
0.01
-0.16
1.00

139

t
4.13
20.36
4.99
-20.63
6.61
-8.61
-1.63
9.28
1.39
-8.22

p-value
0.00
0.00
0.00
0.00
0.00
0.00
0.10
0.00
0.17
0.00

Numerous differences were noted in Table 6.8 and intermittently, I will highlight some
differences between Tables 6.5 and 6.8 because they represent two different spatial aspects of the
Somali population – determinants that reflect where they were (Table 6.5) and determinants that
reflect where they wanted to be as indicated with the migrations (Table 6.8). Many of the
variables associated with the ―Residential Development‖ dimension are down in the order of the
coefficients indicating that they hold less importance for the migrants. The new foreign-born
variable, however, moved up in its importance to the number three position in the coefficient
ordering – this could be explained by the attractiveness of the growing enclaves. The fourth
most influential variable – the climate variable – has also displayed a marked change. Not only
has it become more important by moving up to a higher position in the table, it now has a
negative association with the Somali migrants. This translates into an avoidance of destinations
where there is a plethora of climatic events causing more than $50,000 in damages. The research
done on Somalis in the United States typically focuses on race and assimilation, thus – to this
researcher‘s knowledge – there has been little or nothing written about the relationship (if any)
between the Somalis and the climate here in the United States. This research has therefore
established a link between the migration Somali and the avoidance of adverse climate conditions.

There are a number of other important differences. First, there was a difference noticed
in the race related variables for percent black and white at the migrant destinations. Both
dropped in the coefficient ordering which indicates less emphasis placed on racial demographics.
In addition, the Somali migrants now exhibit a positive association with the percent of the
population white.

140

Table 6.8: PCR results - migration determinants.
Order

Variable

Reconstituted
Coefficient
0.309
0.198
0.172

1
2
3

Somali population
Clan diversity
Percent of the population foreign-born entering between 1990
and 2000

4

Number of climate events causing more than $50,000 in
damages.

-0.125

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

Population density (square mile)
Average utility allowance among households who have it
Number of open units under contract for federal subsidy
Crime per 1000 persons
Total population
Average months on waiting list among admissions
The unemployment rate for the county
Percent of households: family
Vacancy rate for rental units
Number of hospitals
TANF Benefits Per Poor Person
Median real estate taxes
Drop-out rate for teens 16-19
FSP Participation Per Poor Person
Number of physicians
FSP Benefits Per Poor Person
Percent of households: 3 + bedrooms
The estimated per capita number of jobs for the county
median age of the population
Percent of persons insured
Infant mortality rate
Percent of the population black
Percent of population voting Democrat
Percent of the population white
Muslim adherence rate
Median rent asked
Percent of the population in poverty
Average wage per job for the county

-0.106
-0.106
-0.105
-0.100
0.086
-0.086
-0.074
-0.072
-0.072
0.068
0.057
-0.057
0.056
0.054
0.053
0.037
-0.031
0.027
0.016
0.013
0.012
0.010
0.006
0.003
0.003
-0.003
-0.002
-0.002

141

This positive association with the white population could be an indication of the increase in
control the Somali now exercise in their destination choice versus the little control they had upon
initial resettlement in the United States. Secondly, an important difference was noticed with the
―Criminal Propensity‖ dimension.

While the migrant‘s destination crime rate increased in

importance, it also became negatively associated with the Somali migrants. In addition, the age
variable – while decreasing in importance – became positively associated with the migrant
destinations. The differences seen in these two variables could serve as an indication that the
Somali are finding older, more stable neighborhoods in their migration. One repeated reason the
Somali secondary migrants gave for leaving Atlanta, Georgia was that Atlanta was not safe and
Lewiston, Maine was.

The final important difference concerns the variables associated with welfare. All of the
welfare variables – TANF and FSP variables – are now clustered in the middle of the table. In
table 6.5, the TANF occupied a position that reflected more importance and the FSP variables
displayed low importance. The increased importance of the FSP variables can be interpreted as a
reflection of what occurs when refugees leave their initial resettlement community – their
benefits do not follow them to their migration destination. The elevation of the importance of
the FSP variables could indicate their value as a replacement for the loss of the resettlement
assistance. While the overall importance of the welfare variables increased, variables associated
with employment decreased. The per capita number of jobs and the average wage variables
displayed extreme decreases in importance for the migrants – but the unemployment variable
stayed relatively the same. The polarization between the employment and welfare variables can

142

be interpreted as lending support to the welfare magnet hypothesis whereby locations with better
welfare benefits were actively sought out.

The results revealed in this portion of the research leads to the conclusion that first and
foremost, Somali migrants were heavily influenced by the presence of other Somali at the
destination spots.

The results also suggest that while the clan variable was an important

determinant, it was the diversity of clans that served as the attraction for the Somali migrants.
This revelation works counter to my hypothesis that the places sought out would be places with
low diversity indicating a low number of clans or no clans. The migration destinations were
associated with job surpluses, low unemployment, and low poverty; but they were also
associated with a healthy welfare system. From the results, the health of the welfare system held
more importance than the employment aspects of the destination. The migrant destinations were
also characterized as being racially diverse, being relatively safe, and having an older population.

Determination of Cities Associated with Specific Clans
The objective of this step was to empirically determine if there were cities in the United
States specific to certain clans. To answer this question and to paint a broader picture of the
Somali secondary migrations, I drew upon qualitative data derived from surveys, interviews, and
participant observation. It was my hypothesis that with the increased dispersion of Somalis
within the United States, some cities / towns were becoming associated with specific clans.

143

Organization Surveys
To determine if there were cities where specific clans dominated, I used web-based
surveys sent to Somali organizations as well as to voluntary agencies. In all, over 254 emails
were sent out inviting the organizations to participate in this research study. Forty seven of these
emails were returned as ―undeliverable‖ and twelve were returned as ―out of office‖
notifications. Thus, potentially one hundred and ninety-five emails were delivered with the
invitation. Out of those potentially delivered, seven organizations responded by participating in
the survey for a response rate of 3.6 percent. The cities in which the responses came from
include: Philadelphia, Grand Rapids, Denver, Chicago, Omaha, New Haven, and Kansas City.
With such a small response rate, generalizations cannot be made about the target population
without the introduction of severe bias due to the lack of representation. Thus, I could not
establish if certain cities were becoming associated with specific Somali clans. The responses of
the organizations that did participate, however, are provided for reference purposes in Table 6.9.
The thing to note from this table is the question pertaining to the self-sufficiency of secondary
migrants and the clan to which secondary migrants belonged. According to the responses, the
majority of the agencies received Somali secondary migrants, but they did not know the specific
group to which these Somalis belonged. In addition, the majority of the organizations that did
respond felt that the Somali secondary migrants they received were not self-sufficient. This is an
issue that was addressed earlier in this dissertation in reference to the Somali secondary
migrations to Lewiston, Columbus, and Emporia. The reception of secondary migrants who are
not self-sufficient creates additional strain on the resettlement agencies they seek out for help
because the agencies are un-prepared for their arrival.

144

Table 6.9: Organization survey results.
Survey Questions

Percent of
Respondents

Does your agency provide services to Somali refugees?
Yes
85.7
No
14.9
If your agency services Somali refugees, which Somali clan/ethnic group do you
believe make up the majority serviced by your agency:
Darood
14.5
Somali Bantu
28.6
Ogadeni (Darood)
14.3
Unknown
42.6
Does your agency receive Somali secondary migrants?
Yes
57.1
No
42.9
If your agency receives Somali secondary migrants, do you know if they are from a
specific Somali clan/ethnic group?
I don't know
71.4
No
14.3
No answer
14.3
If your agency provides services for Somali secondary migrants, are they selfsufficient when they arrive?
No
57.1
Maybe
14.3
No Answer
28.6
If your agency provides services for Somali refugees, how would you characterize
them?
No different to resettle than other refugees.
14.5
Harder to resettle than other refugees.
42.6
Easier to resettle than other refugees.
14.3
Resettlement varies according to the Somali clan/ethnic group.
28.6

145

Somali Migrant Surveys
The issue of clan appeared to elicit mixed feelings among those who participated in the
survey – some were hesitant to answer the question, others were extremely assertive about
answering the question. Another issue encountered with the survey process was getting subjects
to take the survey. While forty-eight surveys were completed, for every one of the surveys taken
by the respondents, there were anywhere between three and seven rejections – many occurring
once the topic and the title page of the survey was read. This reaction was expected however,
given the sensitive nature of the topic and my role as an outsider asking for information
concerning the topic. The statistical summary for those that participated in the survey are listed
in Table 6.10.

The first round of questions from this survey is listed in Table 6.11. This table contains
the responses to questions concerning the reasons why the subjects decided to migrate to
Columbus. The answers where arranged in multiple choice format. The first question, and the
most important, was why did the subjects decide to migrate to Columbus. The answers to this
question indicated that there are two primary reasons why the subjects decided to migrate to
Columbus. The most prevalent reason for the migration was to start a job (31.3 percent); the
second reason was to find a job (27.1 percent). Thus the motivation was economic. While these
answers may seem similar, they are contingent on two different concepts. Moving to start a new
job in a new city infers security at the end of the move – moving to find a job holds no guarantee
of financial support suggesting some other support system may be available to the migrant.

146

Table 6.10: Somali respondent background (n=48).
Country of
origin:
Ethiopia
Kenya
Somalia
Sex:
Male
Female
Age bracket:
18 - 24
25 - 29
30 - 34
35 - 39
40 - 44

Percent of
Respondents
06.3
10.4
83.3
62.5
37.5
14.6
39.6
35.4
08.3
02.1

Marital status:
Married
Single
Widowed
Clan Family:
Darood
Hawiye
Isaaq
Rahanweyn
None
Other
No answer
Secondary migrant
Yes
No

Percent of
Respondents
35.4
62.5
02.1
29.2
14.6
08.3
04.2
16.7
25.0
02.1
89.6
10.4

Table 6.11: Reason for migrating.
Survey Question:
If you migrated from
another city, please
choose from the items
below, the one that most
applies as a reason for
your migration:

Answer:
To start a job
To find a job
To join immediate family
To join fellow kinsmen
Other
No Answer
Total
Would religion ever
Yes
serve (or has served) as a No
primary reason for you to No Answer
migrate?
Total
Would kinship / clan
Yes
ever serve (or has served) No
as a primary reason for
No Answer
you to migrate?
Total

147

Count
15
13
4
3
8
5
48
1
46
1
48
12
35
1
48

Percent
31.3
27.1
08.3
06.3
16.7
10.4
100.0
02.1
95.8
02.1
100.0
25.0
72.9
02.1
100.0

Support systems like family and kinsmen were included in the answer choices, however, less
than 9 percent of the respondents on each account chose these systems as their reason for
migrating.

The second question of major importance from this grouping involved the influence of
religion as a reason to migrate – in 2006 it was estimated that nearly all of the Somali were Sunni
Muslim (Akiwumi and Estaville 2009). According to 95.8 percent of the respondents from my
survey, religion would never serve as a primary reason to migrate. The response might seem
shocking given the major role that religion plays in the life of Somalis – both home and abroad;
however, this response seems credible when combined with information received during an
interview with one of the gatekeepers in the Somali community. During this interview, the
question arose about the role of religion in the decision for Somalis to embark on secondary
migrations – the interviewee responded by stating that religion would never serve as a primary
reason for secondary migrations because Somalis would and could prey no matter where they are
located, thus religion would never be a primary reason to migrate (A. Mohamed 2010).

The third and crucial question from this group concerned extracting information as to
whether or not clan would serve as a primary reason to migrate. At the root of my hypotheses
are assertions that the migration of the Somali here in the United States was based on clan
movements. Thus, establishing that Somali migrants embarked on secondary migrations to
reconnect with fellow clansmen was one of the goals.

148

According to the responses on this question, nearly 73 percent of the respondents indicated that
clan would never serve as a primary reason to migrate. There were some that declined to answer
the question; however, 25 percent of the respondents did indicate that clan would serve as a
primary motivator for migration. Those that indicated clan would serve as a primary reason did
make comments which I recorded in writing on the survey forms. Many seemed to indicate that
the choice was based on the connections and resources that could be provided by fellow clan
members. Thus it was more a matter of social networking.

The next group of questions (Table 6.12) pertained to the strength several variables had –
or would have – on the decision to migrate. Those variables that would serve as ―very strong‖
reasons to migrate were a better job and better housing. Immediate family, clan or kinsmen, and
better social services, on the other hand, would serve as ―strong‖ influences on the decision to
migrate. While the distinction between ―strong‖ and ―very strong‖ may seem trivial, the fact that
there were substantial gaps for several questions between the percentage of respondents that
chose one or the other shows that the distinction holds significant meaning for them. One area in
which this distinction was very close to being similar, however, concerned the existence of better
social services at the destination spot. According to the respondents, 41.7 percent indicated that
better social services would serve as a ―very strong‖ influence versus 45.8 percent indicating that
it would be a ―strong‖ influence. The question of religion was also incorporated into this group
of questions. The responses for the religion question confirm what was discovered in the initial
group of questions – religion would not serve as a reason to migrate. Nearly 88 percent of the
respondents indicated that religion had – or would have – no strength in influencing their
decision.

149

Table 6.12: Influences on the decision to migrate.
Please indicate the strength of influence the following
variables would (or have had) on your decision to migrate:
Choice
Better Job
Better Housing

Kinsman / Clan

Religion

Immediate Family

Better Social Service

Answer
Strong
Very strong
None
Strong
Very strong
None
Strong
Very strong
None
Strong
No Answer
None
Strong
Very strong
No Answer
None
Strong
Very strong

Frequency Percent
12
25.0
36
75.0
3
6.3
17
35.4
28
58.3
20
41.7
25
52.1
3
6.3
42
87.5
4
8.3
2
4.2
4
8.3
30
62.5
13
27.1
1
2.1
6
12.5
22
45.8
20
41.7

The final group of questions (Table 6.13) was incorporated to get a better sense of the
respondent‘s migration support system or possible constraints. The majority of the migrants
indicated that immediate family (39.6 percent) would be the primary source of economic and
social support at the migration destination. Close to 92 percent indicated that they had relative
living in the same city – so this accounts for the choice of relatives as the primary support givers.
Many (33.3 percent) also indicated ―Other‖ as a support, but the respondents did not elaborate on
what or who this ―Other‖ was. The choices were close to being split on the use of the
government and the use of clan or fellow kinsmen as a support – 12.5 and 10.4 percent
150

respectively. The low figures are understandable in the case of the government – when refugees
move from their initial areas of resettlement, the resettlement services provided by agencies do
not follow them to their new destination in secondary migrations. The next questions concerned
possible barriers to migration. This was a question of whether or not the dominance of another
clan in a potential migration destination would prevent migration to that city.

Table 6.13: Support in migration.
Question
If you have migrated from, or intend to
migrate to another city, who would you
or have you depended on for economic
and social support?
Would you ever avoid moving to a city
if you knew that a Somali clan/ethnic
group (or any ethnic group) you are in
conflict with have a large population
there?
I had relatives already living in the U.S.
when I arrived.

Answer
Immediate family
U.S. Government
Kinsmen/clan
Other
No Answer
Yes
No
No Answer

False
True
False
I have relatives residing in the same city.
True
False
I experienced discrimination by white
True
Americans.
No Answer
False
I experienced discrimination by African True
Americans.
No Answer

151

Count Percent
19
39.6
6
12.5
5
10.4
16
33.3
2
4.2
30
62.5
17
35.4
1
2.1
16
32
4
44
6
41
1
14
33
1

33.3
66.7
8.3
91.7
12.5
85.4
2.1
29.2
68.8
2.1

The vast majority (62.5 percent) indicated that it would serve as a deterrent. Again, comments
were made by the respondents which had bearing on the question – many stated that the presence
of those they were in conflict would not be a deterrent if they had adequate representation from
their own kinsmen or clan. The final two questions from this group concerned the Somalis
experience with racism. The respondents indicated that they experienced discrimination more
from whites than from blacks. According to the survey, 85.4 percent indicated they experienced
discrimination by white Americans versus 68.8 percent noting discrimination from black
Americans. This question, like the other about conflicting clans, was incorporated as a way of
identifying barriers to Somali migration.

The information extracted from the surveys provided valuable insight into the factors
affecting Somali migration. Despite the surveys being conducted solely on those migrants in
Columbus, the results could possibly apply to Somali secondary migrants in other cities. The
reason for the use of the term ―could‖ is because the level of representation is unknown. At the
time of the surveys, during an interview with one of the gatekeepers, the interviewee commented
that close to 95 percent of the Somalis in Columbus were secondary migrants (A. Mohamed
2010). However, the statement could not be backed up by any known database pertaining to
either the number of Somali let alone the number of Somali secondary migrants in Columbus.
Thus, with an unknown population size, in addition to the subjects being a hidden population, the
uncertainty of representation is a constant. Therefore, the results cannot be generalized to the
entire Somali secondary migrant population.

152

CHAPTER 7: SUMMARY AND CONCLUSION

The goal of this research was to gain insight into the Somali refugee population in the
United States by examining several research questions. The first research question pertained to
the determination of whether or not the cities in which the Somali were initially resettled
diversified with respect to the different clan-families. The results from the first test revealed the
answer to be yes – the Simpson Index of Diversity applied to the 1990 to 2009 admission period
yielded an average index of 0.578 which indicated a high level of diversity.

This result

confirmed my first hypothesis that the resettlement cities were highly diversified. I believed that
it was this diversification that would lead the Somalis to disperse and regroup in frontier cities
and towns based on clan membership. Thus, my second research question was the determination
of Somali dispersion from their sites of initial resettlement.

My second hypothesis was

confirmed with the positive results of the relative clustering index which indicated that the
Somali were more clustered in the locations they were sent to than the locations they were
actually recorded as being at. The third step of the research and the third research question
involved the determination of factors associated with the distribution of the Somali across the
United States. It was found that the variables associated with the Somali distribution from the
PCR were remarkably similar to the ideal variables that characterize official resettlement cities.
As it was suggested, the most plausible explanation for the similarities between the Somali
distribution and the resettlement city characteristics was that – the date range for the data
coincides with a period of time in which the Somali would have been considered ―new‖ to this
country. This would have been a time of enclave building in places such as Columbus, OH and
Minneapolis, MN. This was also at a time prior to the entry of the bulk of the Somali population.
Thus the conclusion was that by the year 2000, most of the Somali population that was present in

153

the United States was still confined – or limited – to resettlement locations established by the
Office of Refugee Resettlement. The Somali that left their initial resettlement communities were
not all destined for new frontier communities as evident from some of the analysis performed in
this dissertation – many embarked on secondary migrations to ethnic clusters that were in the
growth process which is seen in the results of the fourth step of this research. One of the
resettlement agency Program Managers from Grand Rapids, MI who responded to the
organization survey indicated that nearly 80 percent of the Somalis that they resettled embarked
on secondary migrations to other cities soon after their arrival. The destination communities for
most were in the Minneapolis area, but a small minority also emigrated to Dallas, TX and
Knoxville, TN.

The fourth step and the fourth research question involved the determination of factors
associated with the Somali migrant destinations. The variables that held the most importance, as
determined from the second PCR, were those that were specific to the Somali – the existence of a
Somali community and the existence of a Somali community characterized by a diverse mixture
of clans-families. I had originally hypothesized that migrations would be to areas with low clan
diversity; this assumption was based on the on-going conflict that has plagued Somalia for the
past two decades. The reasoning behind this assumption was the idea that the conflict between
the clans would be transplanted to the countries of resettlement along with the refugees;
.however, the opposite appears to be the case. In an American context, this lack of conflict –
whether it is an illusion or not – can be explained by observations made by Darboe (2003).
According to Darboe, clan tensions have lessened because Somalis are faced with more pressing
social issues such as learning the English language, and working and raising children in a new

154

environment. Darboe called this collusion between clans ―ethnogenesis‖ and suggested it was
happening because Somalis of different clans live, work, and worship together (p.469). This
notion was also reinforced in my interview with the resettlement staff worker in Columbus.
During the interview, the question was raised concerning the presence of a variety of clans
within the city and if the mixing rekindled the kind of hatred that was witnessed in Somalia. The
interviewee indicated that ―there is still animosity;‖ however, the animosity is not on the same
level as it was in Somalia. She too went on to indicate that the ―tribal groups here‖ have too
much ―other things‖ to worry about than old rivalries (Staff 2010).

While the destination communities were somewhat similar to those where the Somali
initially resided on most aspects, subtle important differences were noted such as less emphasis
on the residential development, an older population, and lower crime rates. Differences were
also seen in the racial demographics – it displayed decreasing importance.

The Somali

destinations were positively associated not only with blacks now, but with whites as well. This
was best described as a reflection on the greater freedom the secondary migrants now exercise
with respect to their migration destinations. Within the resettlement process, refugees have a
minor amount of control over where they are sent with respect to their initial resettlement
community. However, once they enter the United States, their outmigration from that community
– because they are not under any obligations (moral or legal) to stay – is a reflection of the
introduction of the freedom of choice into their decision making process with respect to potential
destinations.

The most plausible explanation for the nature and magnitude of their positive

association with African Americans in their initial communities rests with the actions of the
VOLAGS. As seen in many cases, once the Somali arrive at their initial resettlement community,

155

they are often settled into areas predominantly occupied by African Americans. This placement
trend was noted to be the case in Columbus. However, in their secondary migrations, the
freedom of choice applies and this choice includes the ability to incorporate racial preferences
into the migration decision. Based on the results of the analysis, this preference appears to be
positively associated with the presence of a higher percentage of whites at the destination
community when compared to the percentage at the originating community.

The last notable change occurred with the overall importance of the welfare variables.
For the Somali migrants, variables associated with welfare increased in importance while
variables associated with employment decreased in importance. This finding therefore lends
support to research investigating the theory of welfare magnets. Again, this change in the
importance of the welfare net at the migrant destinations could be linked to the loss of financial
as well as other government benefits that occur when refugees leave their community of initial
resettlement. However, the fact that the welfare variables were increasing in importance while at
the same time the employment variables were decreasing serves to indicate that there is
something else going on besides the loss of benefits – something that warrant more study.

The fifth research question involved the determination of cities associated with specific
clans. This aspect of the research went unanswered because of the extremely low response rate
to the survey invitations that went out to resettlement organizations across the country. The
responses that were received were highly informative, but because of the sample size being nonrepresentative, no generalization could be made from the results.

However, the surveys

conducted in Columbus on the Somali migrants provided sufficient material to answer the

156

remaining research questions.

First, there was the research question pertaining to the

determination of whether or not clan dominance in certain areas could affect the flow of
migration.

The answer to this question is yes. Even though this research was unable to

definitively establish the existence of U.S. cities where specific clans dominate, from the surveys
of the Somali migrants in Columbus, I was able to establish possible reactions if these cities
existed in theory. To the question ―Would you ever avoid moving to a city if you knew that a
Somali clan/ethnic group (or any ethnic group) you are in conflict with have a large population
there?‖ the Somali respondents (62.5 percent) overwhelmingly indicated that it would prevent
their migration to that destination. Subsequent comments made by some of the respondents,
however, revealed that if they had adequate representation from their own clan in that region,
then it would not prevent their migration.

The remaining research question pertained to the determination of whether or not clan is
related to Somali migrations more so than other socio-demographic variables.

From this

research, indicators point to an answer of yes – clan does have more influence than other sociodemographic factors.

The association, however, is more involved than I had originally

conceptualized. In Table 6.6, the clan variable was recorded in the first dimension which also
happened to be the dimension that accounted for the greatest amount of variation in the
migration data. Secondly, the clan variable obtained a position in Table 6.8 that indicated it
exerted the second greatest amount of influence on Somali migration. The variable was based on
diversity, thus looking at the bigger picture, the result can be interpreted in two different ways:
(1) it was not the presence of a single clan, but rather the existence of a multitude of clans that

157

served as a lure for migrants; or (2) it was not the presences of a multitude of clans, but rather the
presence of the migrant‘s clan in the multitude of clans that served as a lure.

The survey conducted in Columbus, however, does provide for a more definitive answer
to the relationship between secondary migrants and clan. While only 6.3 percent of those
surveyed indicated that joining fellow clansmen served as their reason for migrating, nearly 73
percent indicated that clan would never serve as a primary reason for migrating. However, 52
percent of the respondents would go on to indicate that clan would have a strong influence on
their decision to migrate. Thus, from the data it is clear that clan would influence the migration
process more so than other variables, but it would not serve as the primary factor for migrating.
Clan in this case would serve more as a resource or an amenity – a support mechanism or social
network that may or may not be in place at the destination spot. This statement if backed up
further by the statement made from my interview with the resettlement agency staff worker:
You have whole families, families which go then into immediate families. From your
immediate family, it goes to sub-clan all the way to tribal. So when we come here, we are
family, the whole tribe is. That is how we stay connected, that is how you help each
other – by tribe. If you come here, your tribe will do all the necessary things that you
need, so tribal is big. You want a better place to live, you bring your whole tribe so you
can be a community – it‘s your family. (Staff 2010)
In conclusion, the purpose of this research was to contribute valuable information to the
growing body of knowledge on Somalis living in the United States. This line of research,
however, did not focus on the common theme of Somali identity. Rather, the focus of this
research was on the spatial distribution of the Somali in this country of resettlement and the
influence that clan (a highly sensitive issue to the Somali) may have on that distribution. As
evident from the summary in the preceding paragraphs, both quantitative and qualitative methods
were applied in order to answer specific research questions pertaining not only to the

158

determinants of where the Somalis were at in the resettlement, but also to the determinants of
where they were going in their secondary migrations and why. The quantitative methods used
found acceptance for hypotheses for research questions 1 and 2. The quantitative methods also
identified the locational and migration determinants sought for research questions 3 and 4.
Research question 5, however, relied on qualitative data, which included surveys, participant
observation and unstructured interviews. This was the data collected from the case study of the
Somalis in Columbus. And while the research based on quantitative analyses for a study area
covering the United States resulted in reliable results, the analysis for research question 5 –
which was based on qualitative methods – yielded results that could not be generalized to the
entire Somali migrant population and thus requires further study.

The Somali have maintained a position in the top ten refugee sending countries for more
than a decade and until the ongoing conflict in Somalia is brought to an end, Somalia is likely to
remain a top contributor to the refugee flow into the United States. Within the context of this
dissertation, three examples were given in which Somali migrations caught the destination
communities by surprise and in need of extra financial assistance to help settle the secondary
migrants. This was just three examples, but this same event has occurred and is continuing to
occure in numerous communities throughout the United States. Thus, it is imperative that the
motivations behind Somali secondary migrations be understood because the ongoing conflict and
the sustained refugee flow is likely to result in a continuation of these secondary migration
events as more and more Somali are admitted. This research yielded non-generalizable results
from the qualitative survey portion designed to elicit information from VOLAGS because of a
low response rate. It is therefore also suggested that further research into the relationship

159

between the Somali and the VOLAGS in the initial resettlement communities be conducted to
examine possible procedural changes that can be made to make the resettlement process more
efficient or more responsive to the needs of the Somali. Further research may lead to findings
that could possibly diminish or even prevent the need for Somalis to embark on secondary
migrations.

160

APPENDIX

161

Table A.1: Partial listing of the Darood clan-family
DAROOD
Geri
Issak Hassan
Abayoonis Hassan,
Omer Hassan
Ika Hume Geri
Marehan
Howrarsame
Reer Xasan
Reer Faarax Ugaas
Wagardhac
Celi
Calidheere
Awrtable
Reer Cilmi Maxamed
Ogaden
Gumcadle
Maxamed Subeer
Abdalla
Abudwaaq
Maqaabul
Jidwaq
Bartire
Absguul
Yabaree
Leelkase
Maalismogge
Muumin Aadan
Mahamed Aadan
Mohamud Ali
Eyeh
Fiqi Ismail
Hassan Idiris
Reer Ali Sheikh
Suhurre
Reer Awsalaat

Harti
Warsangali
Aadan Siciid
Reer Saalax
Reer Muuse
Reer Qaasim
Adan Yaqub
Ahmed Dhegawayn
Bah Habar Osmaan
Bah Habar Hasan
Bah Idoor
Bah Ogayslabe
Bah Yabare
Bihina Guuleed
Aamir Of Garad
Mohamoud
Cawramale
Colmarabe
Dubeys
Garad Liban (Tuure)
Garwayne
Gobyawuud
Ahmed Omar
Hinjiye
Idamoge
Jibrell Saiid
Muhumud
Ogayslabe
Raage Ali
Reer Omar
Reer Faatax
Reer Garaad
Reer Mohamed
Reer Saalah
Reer Haaji
Reer Yaasuf of Mohamud
Omar

(Harti con't)
Reer Yusuf of Garad
Mohamoud
Riighaye
Waqadsiinye
Warlabe
Husein Iise
Dhulbahante
Xuseen Saciid
Muuse Saciid
Axmed Saciid
Dashiishe
Reer Saleebaan
Reer Ugaas
Reer Boqor
Reer Sakariye
Reer Xaaji
Reer Fahiye
Reer Isxaaq
Tiinle
Kabtaanle
Magalabe
Majeerteen
Wabeeneeye
Tabale
Siwaaqroon
Reer Cumar
Saleebaan
Calì
Ismail
Musa
Ugaar
Abdirahem
Reer Mahmoud
Reer Bicidyahan
Jibraahiil
Calì
Nuux

162

Table A.2: Partial listing of Dir, Isaaq, Hawiye, and Rahanweyn clan-families.
DIR
Madahweyne, Dir
Gariir
Gurgure
Guure
Barsuuq and Madigaan
Layiile
Akisho
Gorejinno
Dayle and Dalbis
Dheere Madaxweyne
Reer Aw Said
Mandaluug, Dir
Gadabursi
Madoobe, Dir
Isse
Mahe Dir
Maxamed Xiniftire
Surre
ISAAQ
Cimraan
Tol Jecle
Reer Hassan
Reer Hussein
Habar Awal
Sacad Muuse
Ciise Muuse
Ayoup
Arap
Garhajis
Eidagale
Habar Yoonis
Habar Jeclo

HAWIYE
Abgaal
Ajuran
Baadicade
Duduble
Fiqishini
Jibade
Garre
Karanle
Kidir
Garjante
Gorgaarte
Gugundhabe
Habar Gidir
Xawadle
Hiilebi
Hintire
Hobay
Iilawaay
Kaariye
Masare
Maxamad Gurgaarte
Mayle Gurgaarte
Moobleen
Murusade
Raaranle
Sheekhaal
Silcis
Saxawle
Ujeejeen
Wacdaan
Wadalaan
Xaskul

163

RAHANWEYN
Digil
Garre
Bagadi
Geledi
Dabarre
Jiido
Shantacaleemood
Tunni
Irroole
Mirifle
Siyeed
Eelaay
Bohoraad
Naasiye
Geedfade
Jiroon
Leysan
Harin
Macalinweyne
Boqolhore
Qoomaal
Disoow
Eemid
Yalaale
Heledi
Haroow
Waanjaal
Reer Dumaal
Garwaale
Sagaal
Hadame
Luwaay
Geelidle
Jilible
Yantaar
Hubeer
Gasaargude
Goobabweyn

Table A.3: Variables for locational determinants.
Aspect
1. Employment

2. Public
housing
3. Public
Assistance
4. Crime
5. Schools

6. Health

7. Climate

8. Population

9. Households

10. Political
11. Religion

Label
wage
jobs
unemp
subunt
wait
utilall
fspart
fsben
tanfben
crm
doutr
hosp
ins
phys
mort
clmt
pop00
pwhite
pblack
pfrbrn
mnage
ppov
pdense
fmlyhh
vac
3bed
mrent
mreal
party

Data Source

muslad

The Association of Religion Data Archives

National Atlas of the United States

Office of Policy Development and Research and HUD USER

US Department of Agriculture: Economic Research Services
National Atlas of the United States
The Annie E. Casey Foundation: Kids Count Data Center
US Department of Health and Human Services
National Atlas of the United States
National Atlas of the United States

American Factfinder

American Factfinder

Election Atlas

164

Table A.4: Variables for migration determinants.
Aspect
1. Employment

2. Public
housing
3. Public
Assistance
4. Crime
5. Schools

6. Health

7. Climate

8. Population

9. Households

10. Political
11. Religion
12. Clan

Label
wage
jobs
unemp
subunt
wait
utilall
fspart
fsben
tanfben
crm
doutr
hosp
ins
phys
mort
clmt
pop00
somcen
pwhite
pblack
pfrbrn
mnage
ppov
pdense
fmlyhh
vac
3bed
mrent
mreal
party
muslad
cland

Data Source
National Atlas of the United States

Office of Policy Development and Research and HUD USER

US Department of Agriculture: Economic Research Services
National Atlas of the United States
The Annie E. Casey Foundation: Kids Count Data Center
US Department of Health and Human Services
National Atlas of the United States
National Atlas of the United States

American Factfinder

American Factfinder

Election Atlas
The Association of Religion Data Archives
WRAPS data

165

Table A.5: Survey for Columbus Somali
1

2

3

4

5

6

7

8

Please list your country and area of origin:
country:
area (city):
Please indicate your sex:
a) Male
b) Female
Please indicate your age bracket:
a) 18 - 24
b) 25 - 29
c) 30 - 34
d) 35 - 39
e) 40 - 44
f) 45+
Please indicate your status:
a) Married
b) Single
c) Divorced
d) Widowed
e) Separated
Do you have any children under the age of 18 that live with you?
a) Yes
b) No
Have you lived in any other U.S.cities prior to living in Columbus?
a) Yes
b) No
If you lived in other cities prior to Columbus, please list each city in order (up to five):
1)
4)
2)
5)
3)
If you migrated from another city, please choose from the items below, the one that most
applies as a reason for your migration:
a) To start a job

166

Table A.5 (cont‘d)

9

10

11

13

14

b) To find a job
c) To join immediate family
d) To join fellow kinsmen
e) Other (please specify):
Please indicate which of the following primary lineage groups you belong to:
a) Daarood
b) Digil
c) Dir
d) Hawiye
e) Isxaaq
f) Rahanwiin
g) None
h) Other (please specify):
Please indicate the strength of influence the following variables would (or have had) on
your decision to migrate:
very strong
strong
none
weak
very weak
a) Better job
b) Better housing
c) Kinsmen / clan
d) Religion
e) Immediate family
f) Better social services
Would religion ever serve (or has served) as a primary reason for you to migrate?
a) Yes
b) No
a) Yes
b) No
If you have migrated from, or intend to migrate to another city, who would you or have
you depended on for economic and social support?
b) U.S. Government
c) Kinsmen/clan
d) Other:
Would you ever avoid moving to a city if you knew that a Somali clan/ethnic
group (or any ethnic group) you are in conflict with have a large population
there?

167

Table A.5 (cont‘d)
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

Year of arrival in the U.S.?
Age at arrival in the U.S.?
First city of residence in the U.S.?
Duration of residence in first city?
Duration of residence in last city?
Aspect of U.S. most attractive?
Principal problem faced since arriving in the U.S.?
Who gave you the most help during your first six months?
I had relatives already living in the U.S. when I arrived.
True
False
I have relatives residing in the same city.
True
False
I have relatives residing in other areas of the U.S.
True
False
I plan to become a U.S. citizen.
True
False
I experienced discrimination by white Americans.
True
False
I experienced discrimination by African Americans.
True
False
I believe there is inequalities in the opportunities in the U.S.
True
False
There is racial discrimination in the U.S.
True
False
American life weakens family
True
False
I have close friends that are not Somali
True
False
If you would like to elaborate or comment on the questions asked, please do so here:

168

Table A.6: Somali Organization Survey Questions
1
2
3

4

5

6

In what year was you organization established?
In what city and state is your organization located?
City:
State:
Does your organization have branches in other cities?
a) Yes
b) No
Does your organization receive U.S. government funding? (federal, state, and/or
local)
a) Yes
b) No
If you answered "No" to the above question, do you feel your organization would
operate better with U.S. government funding?
a) Yes
b) No
Does your organization serve groups other than Somali?
a) Yes
b) No
If you answered "Yes" to the above question, can you list up to three other groups?

7

8

9

1)
2)
3)
Does your organization serve Somalis that have migrated to your area from other
cities within the U.S.?
a) Yes
b) No
Are there multiple Somali clan/ethnic groups in your area?
a) Yes
b) No
c) I don't know

169

Table A.6 (cont‘d)

10

11

If there are multiple Somali clan/ethnic groups in your area, can you list up to four
by their primary lineage group?
1)
2)
3)
4)
5) Don't know
If there are multiple Somali clan/ethnic groups in your area, do they each have their
own organization?
a) Yes
b) No
c) Don't know
Is your organization primarily geared towards service and support for a specific
Somali clan/ethnic group?

12

a) Yes
b) No
c) It depends (please explain):
Does your organization help Somalis regardless of their clan/ethnic group?

13

14

15

16

a) Yes
b) No
Is your organization representative of the largest Somali clan/ethnic group in your
area?
a) Yes
b) No
c) Don't know
Has your organization witness a lot of movement amongst the Somalis that you
service? (i.e. see a lot of the Somalis moving into and out of your area)
a) Yes
b) No
If you answered "Yes" to the above question, can you list the top 3 reasons, in order
of importance, for their movement?
1)
2)
3)
4) Not sure

170

Table A.6 (cont‘d)

17

If you noticed a lot of movement, can you list up to five U.S. cities the Somali
migrants came from?
1)
2)
3)
4)
5)

18

If you have any other comments or would like elaborate on any of the questions
asked or want to share some insight on Somali migration practices; please feel free
to use the space below to voice your opinion:

171

BIBLIOGRAPHY

172

BIBLIOGRAPHY
Abdi, Mohamed, telephone interview by Jay L. Newberry. (June 2010).
Abdullahi, Mohamed. Culture and Customs of Somalia. West Port: Greenwood Press, 2001.
Adam, Hussein. From Tyranny to Anarchy: The Somali Experience. Trenton, NJ: The Red Sea
Press, 2008.
Advameg. Culture of Somalia. 2011. http://www.everyculture.com/Sa-Th/Somalia.html
(accessed June 15, 2011).
Akiwumi, Fenda A., and Lawrence E. Estaville. "Somalis in Maine." In The African Diaspora in
the United States and Canada at the Dawn of the 21st Century, edited by John W.
Frazier, Joe T. Darden and Norah F. Henry, 257-270. Binghamton, NY: Global Academic
Publishing, 2009.
Alexander, Natasha, ed. Somalia: Country Study Guide. Washington, DC: International Buisness
Publications, 2001.
Allard, Scott W., and Sheldon Danziger. "Welfare Magnets: Myth or Reality?" The Journal of
Politics 62, no. 4 (2000): 350-368.
Anderson, Elliot A. "Pirate's Life for Some: The Development of an Illegal Industry in Response
to an Unjust Global Power Dynamic." Indiana Journal of Global Legal Studies 17, no. 2
(2010): 319-339.
Babbie, Earl. The Practice of Social Research. Belmont: Wadworth Publishing Company, 1989.
Barnett, Don. Fraud-Ridden Refugee Program May Restart. September 24, 2010.
http://www.cis.org/barnett/refugee-family-reunification (accessed December 7, 2010).
—. "Out of Africa: Somali Bantu and the Paradigm Shift in Refugee Resettlement." Center for
Immigration Studies. 2003. http://www.cis.org/articles/2003/back1303.pdf (accessed June
7, 2010).
—. Show Me the Money: How Government Funding Has Corrupted Refugee Resettlement. April
1999. http://www.cis.org/print/GovernmentFundingRefugeeResettlement (accessed May
15, 2011).
Bartel, Ann P. "Where Do the New U.S. Immigrants Live?" Journal of Labor and Economics 7
(1989): 371-391.
BBC News. Pirates Seized a Record 1,181 Hostages in 2010 - Report. January 18, 2011.
http://www.bbc.co.uk/news/world-africa-12214905 (accessed May 11, 2011).

173

Berlin, Joey. Somali Farewell. February 29, 2008. http://www.emporiagazette.com/news/2008
/feb/29/somali_farewell/ (accessed October 31, 2009).
Besteman, Catherine. Unraveling Somalia. Philadelphia: University of Pennsylvania Press, 1999.
Biernacki, Patrick, and Dan Waldorf. "Snowball Sampling: Problems and Techniques of Chain
Referral Sampling." Sociological Methods & Research 10 (November 1981): 141-163.
Bjork, Stephanie R. "Modernity Meets Clan: Cultural Intimacy in the Somali Diaspora." In From
Mogadishu to Dixon: The Somali Diaspora in a Global Context, edited by Abdi M.
Kusow and Stephanie R. Bjork, 135-157. Trenton, NJ: Red Sea Press, 2007.
Black, Richard. "Fifty Years of Refugee Studies: From Theory to Policy." International
Migration Review 35 (2001): 57-78.
Borjas, George J. "Welfare Reform and Immigrant Participation in Welfare Programs." In Host
Societies and the Reception of Immigrants, edited by Jeffrey Reitz, 289-325. San Diego,
CA: Center for Comparative Immigration Studies, 2003.
Bouchard, Kelley. Lewiston's Somali Surge. April 28, 2002. http://pressherald.mainetoday.com/
news/immigration/020428lewiston.shtml (accessed January 1, 2009).
Brons, Maria H. Society, Security, Sovereignty and the State in Somalia: From Statelessness to
Statelessness? Utretch: International Books, 2001.
Bruno, Andorra. "Congressional Research Services: Refugee Admissions and Resettlement
Policy." WikiLeaks Document Release. February 2, 2009. http://wikileaks.org/wiki/CRSRL31269 (accessed December 27, 2010).
Bruno, Andorra. Refugee Admissions and Resettlement Policy. CRS Report for Congress,
Washington, D.C.: Congressional Research Service, 2008.
Bryman, Alan. ―Integrating Quantitative and Qualitative Research: How is it Done?‖ Qualitative
Research , 2006: 97-113.
Buckley, F.H. ―The Political Economy of Immigration Policies.‖ International Review of Law
and Economics 16 (1996): 81-99.
Bureau of Population, Refugees, and Migration. "Fraud in the Refugee Family Reunification
(Priority Three) Program." Fact Sheet. U.S. Department of State, February 3, 2009.
Businessdictionary.com. 2010. http://www.businessdictionary.com/definition/economicman.html (accessed December 27, 2010).
Cadwallader, Martin. Migration and Residential Mobility: Macro and Micro Approaches.
Madison: University of Wisconsin Press, 1992.

174

California Department of Social Services. "Voluntary Resettlement Agencies (VOLAGs)."
CDSS. January 2008. http://www.cdss.ca.gov/refugeeprogram/PG1529.htm (accessed
April 23, 2011).
Church World Services. Newsfeature: Secondary Migration to North Texas: "Of course we
help!". July 2, 2009. http://www.churchworldservice.org/site/News2?id=7415
&printer_friendly=1 (accessed June 6, 2011).
Clark, W.A.V. Human Migration. Vol. 7 Scientific Geography Series. Beverly Hills, CA: Sage
Publications, 1986.
CNN Money. Best Places to Live. 2011. http://www.money.cnn.com/magazines/moneymag/
bplive/2006/snapshots/PL3918000.ht ml (accessed May 30, 2011).
Coen, Christopher. Increase in R&P Public Funding - An Analysis. January 31, 2010.
http://www.forefugees.com/2010/01/31/increase-in-rp-public-funding-an-analysis/
(accessed May 7, 2011).
Cohen, Rina, and Gerald Gold. "Constructing Ethnicity: Myth of Return and Modes of Exclusion
among Israelis in Toronto." International Migration, 1997: 373-394.
Community Research Partners. Counting the Franklin County Somali Population. Data Byte No.
2 Somali Population, Columbus: Community Research Partners, 2009, 1-8.
Danso, Ransford. ―From 'There' to 'Here': An Investigation of the Initial Settlement Experiences
of Ethiopian and Somali Refugees in Toronto.‖ GeoJournal 55 (2001): 3-14.
Darboe, Kebba. ―New Immigrants in Minnesota: The Somali Immigration and Assimilation.‖
Journal of Developing Societies 19 (2003): 458-472.
Darden, Joe T., Joshua G. Bagaka‘s, and Shun-Jie Ji. ―Racial Residential Segregation and the
Concentration of Low- and High-Income Households in the 45 Largest U.S. Metropolitan
Areas.‖ Journal of Developing Societies 13 (1997): 171-194.
Davis, Martha F. ―The Evolving Right to Travel: Saenz v. Roe.‖ Publius: The Journal of
Federalism 29, no. 2 (1999): 99-110.
DeJong, T.M. ―A Comparison of Three Diversity Indices Based on Their Components of
Richness and Eveness.‖ Oikos 26, no. 2 (1975): 222-227.
Desbarats, Jacqueline. ―Indochinese Resettlement in the United States.‖ Annals of the
Association of American Geographers 75, no. 4 (1985): 522-538.
Eno, Mohamed A., and Omar A. Eno. ―Intellectualism amid Ethnocentrism: Mukhtar and the 4.5
Factor.‖ Bildaan: An International Journal of Somali Studies 9 (2009): 137-145.

175

Fass, Simon M. ―Through a Glass Darkly: Cause and Effect in Refugee Resettement Policies.‖
Journal of Policy Analysis and Management 4, no. 4 (1985): 554-572.
Fekedulegn, B Desta, J. J Colbert, Jr. R. R Hicks, and Michael E. Schuckers. Coping with
Multicollinearity: An Example on Application of Principal Components Regression in
Dendroecology. NE-721, Newton Square: USDA Forest Service, 2002.
FILES: 2005-2009 American Community Survey Ohio / prepared by the U.S. Census Bureau.
2011.
FILES: Census 2000 Special Tabulation (STP-159) - United States / prepared by the U.S. Census
Bureau. 2002.
Frank, Ove, and Tom Snijders. ―Estimating the Size of Hidden Populations Using Snowball
Sampling.‖ Journal of Official Statistics 10, no. 1 (1994): 53-67.
Frazier, John W. ―Race, Ethnicity, and Place in a Changing America: A Perspective.‖ In Race,
Ethnicity, and Place in a Changing America, by John W. Frazier and Eugene L. TetteyFio, 1-24. Binghamton: Global Academic Publishing, 2006.
Frazier, John W., and Florence M. Margai. ―Culture, Cultural Landscapes and the Historical
Context for Economic Expansion, Immigration and Group Settlements in the United
States.‖ In Multicultural Geographies: The Changing Racial/Ethnic Patterns of the
United States, edited by John W. Frazier and Florence M. Margai, 11-28. Binghamton
University, NY: Global Academic Publishing, 2003.
Gartenstein-Ross, Daveed. Jihad Recruiting Effort May Explain Missing Somalis in Manneapolis
Area. December 4, 2008. http://www.foxnews.com/printer_friendly_story/0,3566,
461865,00.html (accessed March 15, 2011).
Getis, Arthur, Judith Getis, Mark Bjelland, and Jerome D. Fellmann. Introduction to Geography.
13th. New York, NY: McGraw Hill, 2011.
Gold, Steven J. ―Migrant Networks: A Summary and Critique of Relational Approaches to
Interntional Migration.‖ In The Blackwell Companion to Social Inequalities, edited by
Mary Romero and Eric Margolis, 257-285. Malden, MA: Blackwell Publishing, 2005.
—. Refugee Communities: A Comparative Field Study. Newbury Park: Sage Publications, 1992.
Goza, Franklin. ―The Somali Presence in the United States: A Socio-Economic and
Demographic Profile.‖ In From Mogadishu to Dixon: The Somali Diaspora in a Global
Context, by A.M Kusow and S. Bjork, 255-274. Trenton: The Red Sea Press, 2007.

176

Grady, Sue C. ―Low Birth Weight and the Contribution of Residential Segregation, New York
City, 2000.‖ In Multicultural Geographies: The Changing Racial/Ethnic Patterns of the
United States, edited by John W. Frazier and Florence M. Margai, 65-75. Binghamton
University, NY: Global Academic Publishing, 2003.
Grbich, Carol. Qualitative Data Analysis: An Introduction. Thousand Oaks: Sage Publications,
2007.
Gurak, Douglas T., and Mary M. Kritz. ―The Interstate Migration of U.S.Immigrants: Individual
and Contextual Determinants.‖ Social Forces 78, no. 3 (2000): 1017-1039.
Hardwick, Susan W., and James E. Meacham. ―Heterolocalism, Networks of Ethnicity, and
Refugee Communities in the Pacific Northwest: The Portland Story.‖ The Professional
Geographer 57, no. 4 (2005): 539-557.
Hassan, Harun, and Cedric Barnes. A Return to Clan-Politics (or Worse) in Southern Somalia?
March 27, 2007. http://www.hornofafrica.ssrc.org/Hassan_Barnes (accessed May 16,
2011).
Hein, Jeremy. Ethnic Origins: The Adaptation of Cambodian and Hmong Refugees in Four
American Cities. New York, NY: Russell Sage Foundation, 2006.
Hein, Jeremy. ―Refugees, Immigrants, and the State.‖ Annual Review of Sociology 19 (1993): 4359.
Hesse, Brian J. ―Introduction: The Myth of 'Somalia'.‖ Journal of Contemporary African Studies
28, no. 3 (2010): 247-259.
Hill, M.O. ―Diversity and Eveness: A Unifying Notation and Its Consequences.‖ Ecology 54, no.
2 (1973): 427-432.
Holly, Ladislav. Anthropological Perspectives on Kinship. Chicago, IL: Pluto Press, 1996.
Horst, Cindy. "The Somali Diaspora in Minneapolis: Expectations and Realities." In From
Mogadishu to Dixon: The Somali Diaspora in a Global Context, by A.M. Kusow and S.
Bjork, 275-294. Trenton, NJ: Red Sea Press, 2007.
—. Transnational Nomads: How Somalis Cope with Refugee Life in the DaDaab Camps of
Kenya. New York: Berghahn Books, 2006.
Immigration and Refugee Board of Canada. "Victims and Vulnerable Groups in Southern
Somalia." May 1, 1995. http://www.unhcr.org/refworld/docid/3ae6a8092.html (accessed
July 24, 2011).
Jacobsen, Karen. ―Factors Influencing the Policy Responses of Host Governments to Mass
Refugee Influxes.‖ International Migration Review 30, no. 3 (1996): 665-678.

177

Jolliffe, I.T. Principal Component Analysis, Second Edition. New York, NY: Springer-Verlag,
2002.
Juliano, Nick. Ohio Homeless Agencies Struggle with Somali Migrant Influx. Associated Press.
June 18, 2005. http://www.kare11.com/news/news_article.aspx?storyid=99980 (accessed
October 11, 2009).
Kalton, Graham, and Dallas W. Anderson. ―Sampling Rare Populations.‖ Journal of the Royal
Statistical Society 149, no. 1 (1986): 65-82.
Kapteijns, Lidwien, and Abukar Arman. ―Educating Immigrant Youth in the United States: An
Exploration of the Somali Case.‖ Bildhaan: An International Journal of Somali Studies 4
(2004): 18-43.
Keely, Charles B. "Effects of the Immigration Act of 1965 on Selected Population
Characteristics of Immigrants to the United States." Demography 8, no. 2 (1971): 157169.
Kennedy, Edward M. ―Refugee Act of 1980.‖ International Migration Review 15, no. 1/2
(1981): 141-156.
Kusow, Abdi. ―Migration and Racial Formations among Somali Immigrants in North America.‖
Journal of Ethnic and Migration Studies 32 (2006): 533-551.
Kusow, Abdi, and Stephanie Bjork. ―The Somali Diaspora in a Global Context.‖ In From
Mogadishu to Dixon: The Somali Diaspora in a Global Context, by A.M. Kusow and S.
Bjork, 1-9. Trenton, NJ: The Red Sea Press, 2007.
Lanphier, Michael C. ―Refugee Resettlement: Models in Action.‖ International Migration
Review 17 (1983): 4-33.
Larose, Daniel T. Data Mining Methods and Models. Hoboken, NJ: John Wiley & Sons, 2006.
Le Sage, Andre. ―Prospects For Al Itihad and Islamist Radicalism in Somalia.‖ Review of
African Political Economy 27, no. 89 (2001): 472-473.
Lewis, I.M. A Pastoral Democracy. New York: Africana Publishing Company, 1982.
—. Blood and Bone: The Call of Kinship in Somali Society. Lawrenceville, NJ: The Red Sea
Press, 1994.
Lewis, I.M. ―Pan-Africanism and Pan Somalism.‖ Journal of Modern African Studies 1, no. 2
(1963): 147-161.

178

Lewis, I.M. ―The Ogaden and the Fragility of Somali Segmentary Nationalism.‖ African Affairs
88, no. 353 (1989): 573-579.
Lewis, I.M. ―Visible and Invisible Differences: The Somali Paradox.‖ Journal of the
International African Institute 74 (2004): 489-515.
Liu, R.X., J. Kuang, Q. Gong, and X.L. Hou. ―Principal Component Regression Analysis with
SPSS.‖ Computer Methods and Programs in Biomedicine 71 (2003): 141-147.
Mack, N., C. Woodsong, K. MacQueen, G. Guest, and E. Namey. Qualitative Research
Methods: A Data Collector’s Field Guide. NC: Family Health International, 2005.
Massey, Douglas S., and Nancy A. Denton. ―The Dimensions of Residential Segregation.‖ Social
Forces 67 (1988): 281-315.
Menkhaus, Ken. ―The Crisis in Somalia: Tragedy in Five Acts.‖ African Affairs 106 (2007): 357390.
Mlynar, Bobbi. Emporia Readies for More Somali Refugees. November 3, 2007.
http://www.emporiagazette.com/news/2007/nov/03/emporia_readies_more_somali_refug
ees/ (accessed October 31, 2009).
Mohamed, Jama. ―Kinship and Contract in Somali Politics.‖ Journal of the International African
Institute 77 (2007): 226-249.
Mohamud, Abdinur S., and Abdi M. Kusow. ―Why Somalia Continues to Remain a Failed
State.‖ African Renaissance 3 (2006): 13-23.
MSNBC. Four American Hostages Killed by Somali Pirates. February 22, 2011.
http://www.msnbc.msn.com/id/41715530/ns/world_news-africa/t/four-americanhostages-killed-somali-pirates/ (accessed May 15, 2011).
Nadeau, P. ―The Somalis of Lewiston: Community impacts of rapid immigrant movement into a
small homogeneous Maine city.‖ Paper presentation at the New Immigrants in Urban
New England workshop. Brown University, Providence, RI.: Center for the Study of Race
and Ethnicity, 2003.
New York Post. FBI Thwarts Attempted Terror Bombing at Portland Christmas Tree Ceremony.
November 27, 2010. http://www.nypost.com/f/print/news/national/feds_somali
_born_teen_plotted_unsuccessful_WFax7jWItq6hIvb9zjLeIM (accessed March 16,
2011).
Newbold, K. Bruce. ―Estimating the Refugee Population from PUMS Data: Issues and
Demographic Implications.‖ Growth and Change 33, no. 3 (2002): 370-389.

179

O‘Brian, D.J., J.W. Lloyd, and J.B. Kaneene. ―A Principal Components Analysis of Factors
Critical to Participation in Veterinary Lifelong Education Programs.‖ Journal of
Veterinary Medical Education 20, no. 3 (1995): 124-133.
Office of Refugee Resettlement. Office of Refugee Resettlement Report to Congress – FY 2006.
Congressional Report, Washington, DC: Department of Health and Human Services,
2006.
Office of Refugee Resettlement. Office of Refugee Resettlement Report to Congress - FY 2008.
Congressional Report, Washington, DC: Department of Health and Human Services,
2008.
—. The Refugee Act. May 21, 2010. http://www.acf.hhs.gov/programs/orr/policy/refact1.htm
(accessed March 15, 2011).
Ohio Homeland Security. A Guide To Somali Culture. May 2010.
http://www.publicsafety.ohio.gov/links/HLS0075.pdf (accessed May 25, 2011).
Ohio Refugee Services. Somali Bantu in Columbus. 2006.
http://jfs.ohio.gov/refugee/somaliBantu_inColumbus.stm (accessed July 16, 2009).
Osman, Abdulahi A. ―The Somali Conflict and the Role of Inequality, Tribalism, and Clanism.‖
African Renaissance 3, no. 5 (2006): 74-82.
Patrick, Erin. The US Refugee Resettlement Program. June 2004.
http://www.migrationinformation.org/Feature/display.cfm?ID=229 (accessed February
18, 2007).
Petersen, Rebecca D., and Avelardo Valdez. ―Using Snowball-Based Methods in Hidden
Populations to Generate a Randomized Community Sample of Gang-Affiliated
Adolescents.‖ Youth Violence and Juvenile Justice 3 (2005): 151-167.
Pigozzi, Bruce W. ―A Heirarchy of Spatial Marginality Through Spatial Filtering.‖ The
Professional Geographer 56 (2004): 460-470.
Portes, Alejandro. ―Economic Sociology and the Sociology of Immigration: A Conceptual
Overview.‖ In The Economic Sociology of Immigration: Essays on Networks, Ethnicity,
and Entrepreneurship, edited by Alejandro Portes, 1-12. New York, NY: Russell Sage
Foundation, 1995.
Powell, Michael. In Main Town, Sudden Diversity and Controversy. The Washington Post.
October 14, 2002. http://www.highbeam.com/doc/1P2-405221.html (accessed July 10,
2011).
Putnam, Diana B., and Mohamood Cabdi Noor. The Somalis: Their History and Culture.
Washington, DC: Center for Applied Linguistics, 1993.

180

Raghuram, Parvati, Leroi Henry, and Joanna Bornat. ―Difference and Distinction? Non-migrant
and Migrant Networks.‖ Sociology 44, no. 4 (2010): 623-641.
Riddell, J. Barry. ―On Structuring a Migration Model.‖ Geographical Analysis 2, no. 4 (1970):
403-409.
Riley, Christopher. ―Somali-African American Conflict in Linden: Postcolonial Realities and the
Implications for Radical Democracy.‖ Thesis, Department of Geography, Ohio State
University, Columbus, 2005, 1-33.
Robison, Lindon J., A. Allen Schmid, and Marcelo E. Siles. ―Is Social Capital Really Capital.‖
Review of Social Economics 60, no. 1 (2002): 1-21.
Schaid, Jessica, and Zoltan Grossman. ―The Somali Diaspora in Small Midwestern
Communities: The Case of Barron, Wisconsin.‖ In From Mogadishu to Dixon: The
Somali Diaspora in a Global Context , by A.M. Kusow and S. Bjork, 295-319. Trenton,
NJ: The Red Sea Press, 2007.
Simich, Laura. ―Negotiating Boundaries of Refugee Resettlement: A Study of Settlement
Patterns and Social Support.‖ The Canadian Review of Sociology and Anthropology 40
(2003): 575-591.
Singer, Audrey, and Jill H. Wilson. ―From ‗There‘ to ‗Here‘: Refugee Resettlement in
Metropolitan America.‖ Metropolitan Policy Program, Living Cities Census Series.
Washington, DC: The Brookings Institution, 2006.
Sjaastad, Larry A. ―The Costs and Returns of Human Migration.‖ The Journal of Political
Economy 70 (1962): 80-93.
Smith, Woollcott, and Frederick J. Grassle. ―Sampling Properties of a Family of Diversity
Measures.‖ Biometrics 33 (1977): 283-292.
Snow, R.E., J.D. Hutcheson, and J.E. Prather. ―Using Reputational Sampling to Identify
Residential Clusters of Minorities Dispersed in a Large Urban Region: Hispanic in
Atlanta, Georgia.‖ Proceedings of the Section on Survey Research Methods, American
Statistician Association. 1981. 101-106.
Solomon, D. J. Conducting Web-Based Surveys. 2001.
http://pareonline.net/getvn.asp?v=7&n=19 (accessed September 16, 2009).
Staff, CRIS Resettlement Councelor, interview by Jay Newberry. Community Refugee and
Immigration Services (June 17, 2010).
Stein, Barry N. ―The Commitment to Refugee Resettlement.‖ Annals of the American Academy
of Political and Social Science 467 (1983): 187-201.

181

Thai, Hung Cam. ―Ethnic Origins: The Adaptation of Cambodian and Hmong Refugees in Four
American Cities.‖ Social Forces 87, no. 1 (2008): 613-615.
The Columbus Dispatch. Ohio Unemployment Rates. 2011.
http://www.dispatch.com/live/content/databases/community/unemployment.html
(accessed May 30, 2011).
The World Factbook. Somalia. Washington, DC: Central Intelligence Agency, 2008.
Tilove, Jonathan. "Somali Migration Transforms Lewiston." Morning Sentinel. Lewiston,
Maine: Kennebec Journal, September 8, 2002.
U.S. Census Bureau. Census 2000 Summary File 3 (SF 3). 2000.
http://factfinder.census.gov/servlet/DTGeoSearchByListServlet?ds_name=DEC_2000_S
F3_U&_lang=en&_ts=328530095368 (accessed June 8, 2009).
—. QT-P13. Ancestry: 2000. 2011. http://www.factfinder.census.gov (accessed May 25, 2011).
Welch, S. ―Sampling by Referral in a Dispersed Population.‖ Public Opinion Quarterly 39, no. 2
(1975): 237-245.
Wilder Research Center. Speaking for Themselves: A Survey of Hispanic, Hmong, Russian, and
Somali Immigrants in Minneapolis-Saint Paul. Research Report, Saint Paul, MN:
Amherst H. Wilder Foundation, 2000.
Wong, T. ―Purposive and Snowball Sampling in the Study of Ethnic and Mainstream
Community Organizations.‖ Paper presented at the Annual Meeting of the Western
Political Science Association. San Diego, CA, 2008.
Wright, Robert G. "Voluntary Agencies and the Resettlement of Refugees." International
Migration Review 15, no. 1/2 (1981): 157-174.
Yoshimura, Daisuke. Clans in Somalia. Report based on a presentation by Dr. Joakim Gundel
during the Country of Origin Information Workshop., Austrian Centre for Country of
Origin & Assylum Research and Documentation, 2009.
Zavodny, Madeline. "Determinants of Recent Immigrants‘ Locational Choices?" International
Migration Review 33 (1999): 1014-1030.
Zucker, Norman L, and Naomi F Zucker. The Guarded Gate - The Reality of American Refugee
Policy. New York, NY: Harcourt Brace Jovanovich, 1987.
Zucker, Norman. "Refugee Resettlement in the United States: Policy and Problems." Annals of
the American Academy of Political and Social Science 467 (1983): 172-186.

182